# tensorplay.ao.quantization API Source: https://www.tensorplay.cn/docs/api/tensorplay.ao.quantization.html ## Functions 23 [#](#api-tensorplay.ao.quantization.convert) ### convert function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.convert.html) ```python tensorplay.ao.quantization.convert(module, mapping=None, inplace=False, remove_qconfig=True) ``` Swap calibrated float modules for their quantized counterparts. Each target class must expose from_float; activation qparams the quantized kernels require are resolved from the observers attached by [prepare()](/docs/generated/tensorplay.ao.quantization.prepare.html#tensorplay.ao.quantization.prepare) and exposed on the float module. [#](#api-tensorplay.ao.quantization.dequantize) ### dequantize function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.dequantize.html) ```python tensorplay.ao.quantization.dequantize() ``` dequantize.self(Tensor self) -> Tensor dequantize.tensors(Tensor[] tensors) -> Tensor[] [#](#api-tensorplay.ao.quantization.fake_quantize_per_channel) ### fake_quantize_per_channel function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.fake_quantize_per_channel.html) ```python tensorplay.ao.quantization.fake_quantize_per_channel(x, scales, zero_points, axis=0, quant_min=-128, quant_max=127) ``` Applies per-channel fake quantization with fixed affine parameters. Gradient passes through where x lies inside its channel’s representable real range [qmin-zp, qmax-zp]*scale, else zero. [#](#api-tensorplay.ao.quantization.fake_quantize_per_tensor) ### fake_quantize_per_tensor function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.fake_quantize_per_tensor.html) ```python tensorplay.ao.quantization.fake_quantize_per_tensor(x, scale, zero_point, quant_min=-128, quant_max=127) ``` Applies fake quantization with fixed affine parameters. [#](#api-tensorplay.ao.quantization.fuse_modules) ### fuse_modules function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.fuse_modules.html) ```python tensorplay.ao.quantization.fuse_modules(model, modules_to_fuse, inplace=False, fuser_func=None) ``` Fuse supported modules in-place within a model. Parameters: - model – the model to fuse. - modules_to_fuse – an iterable of module-name groups, each a list of fully-qualified attribute paths, e.g. [["conv1", "bn1", "relu1"], ["fc"]]. Groups of length one are ignored (kept for call-site symmetry). - inplace – mutate the model instead of returning a copy. - fuser_func – replacement for fuse_known_modules(). Returns: The fused model (the same object when inplace). [#](#api-tensorplay.ao.quantization.get_observer_state_dict) ### get_observer_state_dict function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.get_observer_state_dict.html) ```python tensorplay.ao.quantization.get_observer_state_dict(model) ``` Collects the calibration state of every observer under model, keyed by module path — the observer counterpart of state_dict(). [#](#api-tensorplay.ao.quantization.int_repr) ### int_repr function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.int_repr.html) ```python tensorplay.ao.quantization.int_repr() ``` int_repr(Tensor self) -> Tensor [#](#api-tensorplay.ao.quantization.is_quantized) ### is_quantized function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.is_quantized.html) ```python tensorplay.ao.quantization.is_quantized() ``` is_quantized(Tensor self) -> bool [#](#api-tensorplay.ao.quantization.load_observer_state_dict) ### load_observer_state_dict function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.load_observer_state_dict.html) ```python tensorplay.ao.quantization.load_observer_state_dict(model, obs_dict) ``` Loads observer stats produced by [get_observer_state_dict()](/docs/generated/tensorplay.ao.quantization.get_observer_state_dict.html#tensorplay.ao.quantization.get_observer_state_dict) back into the matching observers. [#](#api-tensorplay.ao.quantization.prepare) ### prepare function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.prepare.html) ```python tensorplay.ao.quantization.prepare(model, inplace=False, qconfig_spec=None) ``` Attach observers to float modules for post-training calibration. [#](#api-tensorplay.ao.quantization.q_per_channel_axis) ### q_per_channel_axis function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.q_per_channel_axis.html) ```python tensorplay.ao.quantization.q_per_channel_axis() ``` q_per_channel_axis(Tensor self) -> int [#](#api-tensorplay.ao.quantization.q_per_channel_scales) ### q_per_channel_scales function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.q_per_channel_scales.html) ```python tensorplay.ao.quantization.q_per_channel_scales() ``` q_per_channel_scales(Tensor self) -> Tensor [#](#api-tensorplay.ao.quantization.q_per_channel_zero_points) ### q_per_channel_zero_points function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.q_per_channel_zero_points.html) ```python tensorplay.ao.quantization.q_per_channel_zero_points() ``` q_per_channel_zero_points(Tensor self) -> Tensor [#](#api-tensorplay.ao.quantization.q_scale) ### q_scale function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.q_scale.html) ```python tensorplay.ao.quantization.q_scale() ``` q_scale(Tensor self) -> float [#](#api-tensorplay.ao.quantization.q_zero_point) ### q_zero_point function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.q_zero_point.html) ```python tensorplay.ao.quantization.q_zero_point() ``` q_zero_point(Tensor self) -> int [#](#api-tensorplay.ao.quantization.qscheme) ### qscheme function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.qscheme.html) ```python tensorplay.ao.quantization.qscheme() ``` qscheme(Tensor self) -> int [#](#api-tensorplay.ao.quantization.quantize_dynamic) ### quantize_dynamic function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.quantize_dynamic.html) ```python tensorplay.ao.quantization.quantize_dynamic(model, qconfig_spec=None, mapping=None, inplace=False) ``` Swap modules for their dynamic-quantization counterparts. No calibration is performed: dynamic modules quantize activations at inference time and carry statically quantized weights. [#](#api-tensorplay.ao.quantization.quantize_per_channel) ### quantize_per_channel function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.quantize_per_channel.html) ```python tensorplay.ao.quantization.quantize_per_channel() ``` quantize_per_channel(Tensor self, Tensor scales, Tensor zero_points, int axis, ScalarType dtype) -> Tensor [#](#api-tensorplay.ao.quantization.quantize_per_tensor_dynamic) ### quantize_per_tensor_dynamic function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.quantize_per_tensor_dynamic.html) ```python tensorplay.ao.quantization.quantize_per_tensor_dynamic() ``` quantize_per_tensor_dynamic(Tensor self, ScalarType dtype, bool reduce_range) -> Tensor [#](#api-tensorplay.ao.quantization.quantize_per_tensor) ### quantize_per_tensor function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.quantize_per_tensor.html) ```python tensorplay.ao.quantization.quantize_per_tensor() ``` quantize_per_tensor(Tensor self, float scale, int zero_point, ScalarType dtype) -> Tensor quantize_per_tensor.tensor_qparams(Tensor self, Tensor scale, Tensor zero_point, ScalarType dtype) -> Tensor quantize_per_tensor.tensors(Tensor[] tensors, Tensor scales, Tensor zero_points, ScalarType dtype) -> Tensor[] [#](#api-tensorplay.ao.quantization.quantize) ### quantize function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.quantize.html) ```python tensorplay.ao.quantization.quantize(model, run_fn, run_args, mapping=None, inplace=False) ``` Prepare, calibrate through run_fn(*run_args) and convert. Only submodules carrying a qconfig are quantized; attach qconfigs (or pass a qconfig_dict through propagate_qconfig_()) before calling. [#](#api-tensorplay.ao.quantization.quantized_linear_dynamic) ### quantized_linear_dynamic function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.quantized_linear_dynamic.html) ```python tensorplay.ao.quantization.quantized_linear_dynamic() ``` quantized_linear_dynamic(Tensor input, Tensor weight, Tensor weight_scales, Tensor weight_zero_points, Tensor? bias=None, bool reduce_range=False) -> Tensor [#](#api-tensorplay.ao.quantization.quantized_linear) ### quantized_linear function[Full reference ↗](/docs/generated/tensorplay.ao.quantization.quantized_linear.html) ```python tensorplay.ao.quantization.quantized_linear() ``` quantized_linear(Tensor input, Tensor weight, float input_scale, int input_zero_point, Tensor weight_scales, Tensor weight_zero_points, Tensor? bias=None, float out_scale=1.0, int out_zero_point=0) -> Tensor ## Classes 13 [#](#api-tensorplay.ao.quantization.DeQuantStub) ### DeQuantStub class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.DeQuantStub.html) ```python class tensorplay.ao.quantization.DeQuantStub(scale=None, zero_point=None) ``` ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.DeQuantStub.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.DeQuantStub.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.DeQuantStub.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.DeQuantStub.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.DeQuantStub.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.DeQuantStub.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.DeQuantStub.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.DeQuantStub.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after forward() has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before forward() is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.DeQuantStub.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.DeQuantStub.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.DeQuantStub.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.DeQuantStub.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.DeQuantStub.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.DeQuantStub.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.FakeQuantize) ### FakeQuantize class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.FakeQuantize.html) ```python class tensorplay.ao.quantization.FakeQuantize(observer=None, scale=None, zero_point=None, disable_observer=False) ``` Calibrating / simulating module. With no qparams set, the first forward pass derives scale/zero_point from its observer over incoming batches; call [freeze()](#tensorplay.ao.quantization.FakeQuantize.freeze) to stop recalibrating. With explicit scale/zero_point arguments it is stateless. ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python freeze() ``` Stops calibration and fixes the current derived qparams. ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.FakeQuantize.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.FakeQuantize.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.FakeQuantize.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.FakeQuantize.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.FakeQuantize.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.FakeQuantize.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.FakeQuantize.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.FakeQuantize.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after forward() has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before forward() is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.FakeQuantize.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.FakeQuantize.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.FakeQuantize.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.FakeQuantize.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.FakeQuantize.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.FakeQuantize.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.FixedQParamsObserver) ### FixedQParamsObserver class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.FixedQParamsObserver.html) ```python class tensorplay.ao.quantization.FixedQParamsObserver(scale, zero_point, dtype=tensorplay.int8, quant_min=-128, quant_max=127) ``` Reports fixed scale/zero_point without observing data; used when the quantization parameters are dictated by construction (sigmoid/tanh style ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python forward(*input: Any) → None ``` Define the computation performed at every call. Should be overridden by all subclasses. > **Note** > > Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them. ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.FixedQParamsObserver.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.FixedQParamsObserver.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.FixedQParamsObserver.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.FixedQParamsObserver.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.FixedQParamsObserver.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.FixedQParamsObserver.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.FixedQParamsObserver.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.FixedQParamsObserver.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after [forward()](#tensorplay.ao.quantization.FixedQParamsObserver.forward) has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after [forward()](#tensorplay.ao.quantization.FixedQParamsObserver.forward) is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before [forward()](#tensorplay.ao.quantization.FixedQParamsObserver.forward) is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.FixedQParamsObserver.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.FixedQParamsObserver.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.FixedQParamsObserver.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.FixedQParamsObserver.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.FixedQParamsObserver.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.FixedQParamsObserver.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.HistogramObserver) ### HistogramObserver class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.HistogramObserver.html) ```python class tensorplay.ao.quantization.HistogramObserver(bins=2048, dtype=tensorplay.int8, quant_min=-128, quant_max=127, eps=None) ``` Running-histogram observer. Records a running histogram of incoming values together with the global min/max; calculate_qparams narrows the range with an L2-quantization- error search (the norm-minimization formulation) before deriving affine parameters, which filters outliers instead of trusting raw extremes. ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python forward(*input: Any) → None ``` Define the computation performed at every call. Should be overridden by all subclasses. > **Note** > > Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them. ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.HistogramObserver.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.HistogramObserver.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.HistogramObserver.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.HistogramObserver.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.HistogramObserver.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.HistogramObserver.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.HistogramObserver.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.HistogramObserver.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after [forward()](#tensorplay.ao.quantization.HistogramObserver.forward) has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after [forward()](#tensorplay.ao.quantization.HistogramObserver.forward) is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before [forward()](#tensorplay.ao.quantization.HistogramObserver.forward) is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.HistogramObserver.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.HistogramObserver.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.HistogramObserver.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.HistogramObserver.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.HistogramObserver.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.HistogramObserver.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.MinMaxObserver) ### MinMaxObserver class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.MinMaxObserver.html) ```python class tensorplay.ao.quantization.MinMaxObserver(dtype=tensorplay.int8, quant_min=-128, quant_max=127, eps=None) ``` Tracks the running min/max of observed tensors; per-tensor params. ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python forward(*input: Any) → None ``` Define the computation performed at every call. Should be overridden by all subclasses. > **Note** > > Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them. ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.MinMaxObserver.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.MinMaxObserver.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.MinMaxObserver.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.MinMaxObserver.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python observation_state() ``` Serializable calibration state (tensors / numbers / None). Observers keep their statistics as plain attributes rather than registered buffers, so state_dict() cannot see them; this pair is what get/load_observer_state_dict persist. ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.MinMaxObserver.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.MinMaxObserver.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.MinMaxObserver.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.MinMaxObserver.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after [forward()](#tensorplay.ao.quantization.MinMaxObserver.forward) has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after [forward()](#tensorplay.ao.quantization.MinMaxObserver.forward) is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before [forward()](#tensorplay.ao.quantization.MinMaxObserver.forward) is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.MinMaxObserver.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.MinMaxObserver.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.MinMaxObserver.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.MinMaxObserver.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.MinMaxObserver.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.MinMaxObserver.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.MovingAverageMinMaxObserver) ### MovingAverageMinMaxObserver class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.MovingAverageMinMaxObserver.html) ```python class tensorplay.ao.quantization.MovingAverageMinMaxObserver(averaging_constant=0.01, dtype=tensorplay.int8, quant_min=-128, quant_max=127, eps=None) ``` Exponential moving average of min/max, as used for QAT-style calibration on streamed data. ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python forward(*input: Any) → None ``` Define the computation performed at every call. Should be overridden by all subclasses. > **Note** > > Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them. ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python observation_state() ``` Serializable calibration state (tensors / numbers / None). Observers keep their statistics as plain attributes rather than registered buffers, so state_dict() cannot see them; this pair is what get/load_observer_state_dict persist. ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after [forward()](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.forward) has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after [forward()](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.forward) is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before [forward()](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.forward) is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.MovingAverageMinMaxObserver.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver) ### MovingAveragePerChannelMinMaxObserver class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.html) ```python class tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver(averaging_constant=0.01, ch_axis=0, dtype=tensorplay.int8, quant_min=-128, quant_max=127, eps=None) ``` Exponential moving average of per-channel min/max values. ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python forward(*input: Any) → None ``` Define the computation performed at every call. Should be overridden by all subclasses. > **Note** > > Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them. ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after [forward()](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.forward) has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after [forward()](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.forward) is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before [forward()](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.forward) is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.MovingAveragePerChannelMinMaxObserver.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.PerChannelFakeQuantize) ### PerChannelFakeQuantize class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.PerChannelFakeQuantize.html) ```python class tensorplay.ao.quantization.PerChannelFakeQuantize(ch_axis=0, observer=None, scales=None, zero_points=None, disable_observer=False) ``` Per-channel fake quantization with range-masked STE. ch_axis selects the quantized dimension; scale/zero_point may be given explicitly (tensors of length n) or derived from a PerChannelMinMaxObserver over incoming batches. ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.PerChannelFakeQuantize.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.PerChannelFakeQuantize.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.PerChannelFakeQuantize.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.PerChannelFakeQuantize.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.PerChannelFakeQuantize.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.PerChannelFakeQuantize.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.PerChannelFakeQuantize.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.PerChannelFakeQuantize.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after forward() has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before forward() is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.PerChannelFakeQuantize.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.PerChannelFakeQuantize.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.PerChannelFakeQuantize.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.PerChannelFakeQuantize.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.PerChannelFakeQuantize.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.PerChannelFakeQuantize.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.PerChannelMinMaxObserver) ### PerChannelMinMaxObserver class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.PerChannelMinMaxObserver.html) ```python class tensorplay.ao.quantization.PerChannelMinMaxObserver(ch_axis=0, dtype=tensorplay.int8, quant_min=-128, quant_max=127, eps=None) ``` Running per-channel min/max along ch_axis; returns per-channel scale/zero_point tensors suitable for quantize_per_channel. ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python forward(*input: Any) → None ``` Define the computation performed at every call. Should be overridden by all subclasses. > **Note** > > Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them. ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.PerChannelMinMaxObserver.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.PerChannelMinMaxObserver.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.PerChannelMinMaxObserver.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.PerChannelMinMaxObserver.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.PerChannelMinMaxObserver.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.PerChannelMinMaxObserver.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.PerChannelMinMaxObserver.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.PerChannelMinMaxObserver.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after [forward()](#tensorplay.ao.quantization.PerChannelMinMaxObserver.forward) has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after [forward()](#tensorplay.ao.quantization.PerChannelMinMaxObserver.forward) is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before [forward()](#tensorplay.ao.quantization.PerChannelMinMaxObserver.forward) is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.PerChannelMinMaxObserver.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.PerChannelMinMaxObserver.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.PerChannelMinMaxObserver.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.PerChannelMinMaxObserver.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.PerChannelMinMaxObserver.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.PerChannelMinMaxObserver.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.PlaceholderObserver) ### PlaceholderObserver class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.PlaceholderObserver.html) ```python class tensorplay.ao.quantization.PlaceholderObserver(dtype=tensorplay.float32, custom_op_name='', quant_min=None, quant_max=None, eps=None) ``` No-op observer that only carries configuration, e.g. for float16 “quantization” or dynamic-quantization markers that need no ranges. ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python forward(*input: Any) → None ``` Define the computation performed at every call. Should be overridden by all subclasses. > **Note** > > Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them. ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.PlaceholderObserver.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.PlaceholderObserver.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.PlaceholderObserver.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.PlaceholderObserver.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python observation_state() ``` Serializable calibration state (tensors / numbers / None). Observers keep their statistics as plain attributes rather than registered buffers, so state_dict() cannot see them; this pair is what get/load_observer_state_dict persist. ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.PlaceholderObserver.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.PlaceholderObserver.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.PlaceholderObserver.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.PlaceholderObserver.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after [forward()](#tensorplay.ao.quantization.PlaceholderObserver.forward) has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after [forward()](#tensorplay.ao.quantization.PlaceholderObserver.forward) is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before [forward()](#tensorplay.ao.quantization.PlaceholderObserver.forward) is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.PlaceholderObserver.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.PlaceholderObserver.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.PlaceholderObserver.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.PlaceholderObserver.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.PlaceholderObserver.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.PlaceholderObserver.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. [#](#api-tensorplay.ao.quantization.QConfig) ### QConfig class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.QConfig.html) ```python class tensorplay.ao.quantization.QConfig(activation, weight) ``` Describes how to quantize a layer: observer factories for activations and weights. Both fields must be observer classes or callables returning observer instances. Use MyObserver.with_args(x=1) to override constructor arguments. ```python activation ``` Alias for field number 0 ```python count(value, /) ``` Return number of occurrences of value. ```python index(value, start=0, stop=9223372036854775807, /) ``` Return first index of value. Raises ValueError if the value is not present. ```python weight ``` Alias for field number 1 [#](#api-tensorplay.ao.quantization.QConfigMapping) ### QConfigMapping class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.QConfigMapping.html) ```python class tensorplay.ao.quantization.QConfigMapping ``` Pattern-based assignment of qconfigs to modules. Resolution order for a module: exact fully-qualified name, then module type, then the global qconfig. [#](#api-tensorplay.ao.quantization.QuantStub) ### QuantStub class[Full reference ↗](/docs/generated/tensorplay.ao.quantization.QuantStub.html) ```python class tensorplay.ao.quantization.QuantStub(qconfig=None) ``` ```python add_module(name: str, module: Module | None) → None ``` Add a child module to the current module. The module can be accessed as an attribute using the given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the child module. The child module can be accessed from this module using the given name - module (Module) – child module to be added to the module. ```python apply(fn: Callable[[Module], None]) → Self ``` Apply fn recursively to every submodule (as returned by .children()) as well as self. Typical use includes initializing the parameters of a model (see also [tensorplay.nn.init](/docs/nn.init.html#nn-init-doc)). Parameters: fn (Module -> None) – function to be applied to each submodule Returns: self Return type: Module Example: ``` >>> @tensorplay.no_grad() >>> def init_weights(m): >>> print(m) >>> if type(m) == nn.Linear: >>> m.weight.fill_(1.0) >>> print(m.weight) >>> net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) >>> net.apply(init_weights) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Linear(in_features=2, out_features=2, bias=True) Parameter containing: tensor([[1., 1.], [1., 1.]], requires_grad=True) Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) ``` ```python bfloat16() → Self ``` Casts all floating point parameters and buffers to bfloat16 datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python buffers(recurse: bool = True) → Iterator[Tensor] ``` Return an iterator over module buffers. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Yields: tensorplay.Tensor – module buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for buf in model.buffers(): >>> print(type(buf), buf.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python children() → Iterator[Module] ``` Return an iterator over immediate children modules. Yields: Module – a child module ```python compile(*args, **kwargs) ``` Compile this Module’s forward using tensorplay.compile(). This Module’s __call__ method is compiled and all arguments are passed as-is to tensorplay.compile(). See tensorplay.compile() for details on the arguments for this function. ```python cpu() → Self ``` Move all model parameters and buffers to the CPU. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python cuda(device: int | device | None = None) → Self ``` Move all model parameters and buffers to the GPU. This also makes associated parameters and buffers different objects. So it should be called before constructing the optimizer if the module will live on GPU while being optimized. > **Note** > > This method modifies the module in-place. Parameters: device ([int](https://docs.python.org/3/builtins/functions.html#int), optional) – if specified, all parameters will be copied to that device Returns: self Return type: Module ```python double() → Self ``` Casts all floating point parameters and buffers to double datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python eval() → Self ``` Set the module in evaluation mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e. whether they are affected, e.g. Dropout, BatchNorm, etc. This is equivalent with self.train(False). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .eval() and several similar mechanisms that may be confused with it. Returns: self Return type: Module ```python extra_repr() → str ``` Return the extra representation of the module. To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable. ```python float() → Self ``` Casts all floating point parameters and buffers to float datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python get_buffer(target: str) → Tensor ``` Return the buffer given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The buffer referenced by target Return type: [tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not a buffer ```python get_extra_state() → Any ``` Return any extra state to include in the module’s state_dict. Implement this and a corresponding [set_extra_state()](#tensorplay.ao.quantization.QuantStub.set_extra_state) for your module if you need to store extra state. This function is called when building the module’s state_dict(). Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes. Returns: Any extra state to store in the module’s state_dict Return type: [object](https://docs.python.org/3/builtins/functions.html#object) ```python get_parameter(target: str) → Parameter ``` Return the parameter given by target if it exists, otherwise throw an error. See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target. Parameters: target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.) Returns: The Parameter referenced by target Return type: tensorplay.nn.Parameter Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If the target string references an invalid path or resolves to something that is not an nn.Parameter ```python get_submodule(target: str) → Module ``` Return the submodule given by target if it exists, otherwise throw an error. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) ) ``` (The diagram shows an nn.Module A. A which has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv"). The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_modules achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists, get_submodule should always be used. Parameters: target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) Returns: The submodule referenced by target Return type: tensorplay.nn.Module Raises: [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python half() → Self ``` Casts all floating point parameters and buffers to half datatype. > **Note** > > This method modifies the module in-place. Returns: self Return type: Module ```python load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False) ``` Copy parameters and buffers from [state_dict](#tensorplay.ao.quantization.QuantStub.state_dict) into this module and its descendants. If strict is True, then the keys of [state_dict](#tensorplay.ao.quantization.QuantStub.state_dict) must exactly match the keys returned by this module’s state_dict() function. > **Warning** > > If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True. Parameters: - state_dict ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – a dict containing parameters and persistent buffers. - strict ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to strictly enforce that the keys in [state_dict](#tensorplay.ao.quantization.QuantStub.state_dict) match the keys returned by this module’s state_dict() function. Default: True - assign ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – When set to False, the properties of the tensors in the current module are preserved whereas setting it to True preserves properties of the Tensors in the state dict. The only exception is the requires_grad field of Parameter for which the value from the module is preserved. Default: False Returns: - missing_keys is a list of str containing any keys that are expectedby this module but missing from the provided state_dict. - unexpected_keys is a list of str containing the keys that are notexpected by this module but present in the provided state_dict. Return type: NamedTuple with missing_keys and unexpected_keys fields > **Note** > > If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError. ```python modules() → Iterator[Module] ``` Return an iterator over all modules in the network. Yields: Module – a module in the network > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.modules()): ... print(idx, '->', m) 0 -> Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) ) 1 -> Linear(in_features=2, out_features=2, bias=True) ``` ```python named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Tensor]] ``` Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all buffer names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module. Defaults to True. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated buffers in the result. Defaults to True. Yields: (str, tensorplay.Tensor) – Tuple containing the name and buffer Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, buf in self.named_buffers(): >>> if name in ['running_var']: >>> print(buf.size()) ``` ```python named_children() → Iterator[tuple[str, Module]] ``` Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself. Yields: (str, Module) – Tuple containing a name and child module Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, module in model.named_children(): >>> if name in ['conv4', 'conv5']: >>> print(module) ``` ```python named_modules(memo: set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True) ``` Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself. Parameters: - memo – a memo to store the set of modules already added to the result - prefix – a prefix that will be added to the name of the module - remove_duplicate – whether to remove the duplicated module instances in the result or not Yields: (str, Module) – Tuple of name and module > **Note** > > Duplicate modules are returned only once. In the following example, l will be returned only once. Example: ``` >>> l = nn.Linear(2, 2) >>> net = nn.Sequential(l, l) >>> for idx, m in enumerate(net.named_modules()): ... print(idx, '->', m) 0 -> ('', Sequential( (0): Linear(in_features=2, out_features=2, bias=True) (1): Linear(in_features=2, out_features=2, bias=True) )) 1 -> ('0', Linear(in_features=2, out_features=2, bias=True)) ``` ```python named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[tuple[str, Parameter]] ``` Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself. Parameters: - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – prefix to prepend to all parameter names. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. - remove_duplicate ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – whether to remove the duplicated parameters in the result. Defaults to True. Yields: (str, Parameter) – Tuple containing the name and parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for name, param in self.named_parameters(): >>> if name in ['bias']: >>> print(param.size()) ``` ```python parameters(recurse: bool = True) → Iterator[Parameter] ``` Return an iterator over module parameters. This is typically passed to an optimizer. Parameters: recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module. Yields: Parameter – module parameter Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> for param in model.parameters(): >>> print(type(param), param.size()) (20L,) (20L, 1L, 5L, 5L) ``` ```python record(x) ``` Calibration entry point: feeds the batch to the inner FakeQuantize observer without fake-quantizing (as manual calibration loops do). ```python register_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None ``` Add a buffer to the module. This is typically used to register a buffer that should not be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s [state_dict](#tensorplay.ao.quantization.QuantStub.state_dict). Buffers can be accessed as attributes using given names. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the buffer. The buffer can be accessed from this module using the given name - tensor ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) or None) – buffer to be registered. If None, then operations that run on buffers, such as [cuda](#tensorplay.ao.quantization.QuantStub.cuda), are ignored. If None, the buffer is not included in the module’s [state_dict](#tensorplay.ao.quantization.QuantStub.state_dict). - persistent ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether the buffer is part of this module’s [state_dict](#tensorplay.ao.quantization.QuantStub.state_dict). Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> self.register_buffer('running_mean', tensorplay.zeros(num_features)) ``` ```python register_forward_hook(hook: Callable[[T, tuple[Any, ...], Any], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle ``` Register a forward hook on the module. The hook will be called every time after forward() has computed an output. If with_kwargs is False or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature: ``` hook(module, args, output) -> None or modified output ``` If with_kwargs is True, the forward hook will be passed the kwargs given to the forward function and be expected to return the output possibly modified. The hook should have the following signature: ``` hook(module, args, kwargs, output) -> None or modified output ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the provided hook will be fired before all existing forward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this tensorplay.nn.Module. Note that global forward hooks registered with register_module_forward_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True, the hook will be passed the kwargs given to the forward function. Default: False - always_call ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If True the hook will be run regardless of whether an exception is raised while calling the Module. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_forward_pre_hook(hook: Callable[[T, tuple[Any, ...]], Any | None] | Callable[[T, tuple[Any, ...], dict[str, Any]], tuple[Any, dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle ``` Register a forward pre-hook on the module. The hook will be called every time before forward() is invoked. If with_kwargs is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to the forward. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature: ``` hook(module, args) -> None or modified input ``` If with_kwargs is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature: ``` hook(module, args, kwargs) -> None or a tuple of modified input and kwargs ``` Parameters: - hook (Callable) – The user defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing forward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this tensorplay.nn.Module. Note that global forward_pre hooks registered with register_module_forward_pre_hook() will fire before all hooks registered by this method. Default: False - with_kwargs ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the hook will be passed the kwargs given to the forward function. Default: False Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward hook on the module. The hook will be called every time the gradients with respect to a module are computed, and its firing rules are as follows: - Ordinarily, the hook fires when the gradients are computed with respect to the module inputs. - If none of the module inputs require gradients, the hook will fire when the gradients are computed with respect to module outputs. - If none of the module outputs require gradients, then the hooks will not fire. The hook should have the following signature: ``` hook(module, grad_input, grad_output) -> tuple(Tensor) or None ``` The grad_input and grad_output are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this tensorplay.nn.Module. Note that global backward hooks registered with register_module_full_backward_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_full_backward_pre_hook(hook: Callable[[Module, tuple[Tensor, ...] | Tensor], None | tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle ``` Register a backward pre-hook on the module. The hook will be called every time the gradients for the module are computed. The hook should have the following signature: ``` hook(module, grad_output) -> tuple[Tensor] or None ``` The grad_output is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place of grad_output in subsequent computations. Entries in grad_output will be None for all non-Tensor arguments. For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function. > **Warning** > > Modifying inputs inplace is not allowed when using backward hooks and will raise an error. Parameters: - hook (Callable) – The user-defined hook to be registered. - prepend ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – If true, the provided hook will be fired before all existing backward_pre hooks on this tensorplay.nn.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this tensorplay.nn.Module. Note that global backward_pre hooks registered with register_module_full_backward_pre_hook() will fire before all hooks registered by this method. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_post_hook(hook) ``` Register a post-hook to be run after module’s load_state_dict() is called. It should have the following signature:: hook(module, incompatible_keys) -> None The module argument is the current module that this hook is registered on, and the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_keys is a list of str containing the unexpected keys. The given incompatible_keys can be modified inplace if needed. Note that the checks performed when calling [load_state_dict()](#tensorplay.ao.quantization.QuantStub.load_state_dict) with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=True, and clearing out both missing and unexpected keys will avoid an error. Returns: a handle that can be used to remove the added hook by calling handle.remove() Return type: tensorplay.utils.hooks.RemovableHandle ```python register_load_state_dict_pre_hook(hook) ``` Register a pre-hook to be run before module’s load_state_dict() is called. It should have the following signature:: hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950 Parameters: hook (Callable) – Callable hook that will be invoked before loading the state dict. ```python register_module(name: str, module: Module | None) → None ``` Alias for [add_module()](#tensorplay.ao.quantization.QuantStub.add_module). ```python register_parameter(name: str, param: Parameter | None) → None ``` Add a parameter to the module. The parameter can be accessed as an attribute using given name. Parameters: - name ([str](https://docs.python.org/3/builtins/stdtypes.html#str)) – name of the parameter. The parameter can be accessed from this module using the given name - param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as [cuda](#tensorplay.ao.quantization.QuantStub.cuda), are ignored. If None, the parameter is not included in the module’s [state_dict](#tensorplay.ao.quantization.QuantStub.state_dict). ```python register_state_dict_post_hook(hook) ``` Register a post-hook for the state_dict() method. It should have the following signature:: hook(module, state_dict, prefix, local_metadata) -> None The registered hooks can modify the state_dict inplace. ```python register_state_dict_pre_hook(hook) ``` Register a pre-hook for the state_dict() method. It should have the following signature:: hook(module, prefix, keep_vars) -> None The registered hooks can be used to perform pre-processing before the state_dict call is made. ```python requires_grad_(requires_grad: bool = True) → Self ``` Change if autograd should record operations on parameters in this module. This method sets the parameters’ requires_grad attributes in-place. This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training). See [Locally disabling gradient computation](/docs/notes/autograd.html#locally-disable-grad-doc) for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it. Parameters: requires_grad ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether autograd should record operations on parameters in this module. Default: True. Returns: self Return type: Module ```python set_extra_state(state: Any) → None ``` Set extra state contained in the loaded state_dict. This function is called from [load_state_dict()](#tensorplay.ao.quantization.QuantStub.load_state_dict) to handle any extra state found within the state_dict. Implement this function and a corresponding [get_extra_state()](#tensorplay.ao.quantization.QuantStub.get_extra_state) for your module if you need to store extra state within its state_dict. Parameters: state ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict)) – Extra state from the state_dict ```python set_submodule(target: str, module: Module, strict: bool = False) → None ``` Set the submodule given by target if it exists, otherwise throw an error. > **Note** > > If strict is set to False (default), the method will replace an existing submodule or create a new submodule if the parent module exists. If strict is set to True, the method will only attempt to replace an existing submodule and throw an error if the submodule does not exist. For example, let’s say you have an nn.Module A that looks like this: ``` A( (net_b): Module( (net_c): Module( (conv): Conv2d(3, 3, 3) ) (linear): Linear(3, 3) ) ) ``` (The diagram shows an nn.Module A. A has a nested submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.) To override the Conv2d with a new submodule Linear, you could call set_submodule("net_b.net_c.conv", nn.Linear(1, 1)) where strict could be True or False To add a new submodule Conv2d to the existing net_b module, you would call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1)). In the above if you set strict=True and call set_submodule("net_b.conv", nn.Conv2d(1, 1, 1), strict=True), an AttributeError will be raised because net_b does not have a submodule named conv. Parameters: - target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.) - module – The module to set the submodule to. - strict – If False, the method will replace an existing submodule or create a new submodule if the parent module exists. If True, the method will only attempt to replace an existing submodule and throw an error if the submodule doesn’t already exist. Raises: - [ValueError](https://docs.python.org/3/builtins/exceptions.html#ValueError) – If the target string is empty or if module is not an instance of nn.Module. - [AttributeError](https://docs.python.org/3/builtins/exceptions.html#AttributeError) – If at any point along the path resulting from the target string the (sub)path resolves to a non-existent attribute name or an object that is not an instance of nn.Module. ```python share_memory() → Self ``` See tensorplay.Tensor.share_memory_(). ```python state_dict(*args, destination=None, prefix='', keep_vars=False) ``` Return a dictionary containing references to the whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Keys are corresponding parameter and buffer names. Parameters and buffers set to None are not included. > **Note** > > The returned object is a shallow copy. It contains references to the module’s parameters and buffers. > **Warning** > > Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases. > **Warning** > > Please avoid the use of argument destination as it is not designed for end-users. Parameters: - destination ([dict](https://docs.python.org/3/builtins/stdtypes.html#dict), optional) – If provided, the state of module will be updated into the dict and the same object is returned. Otherwise, an OrderedDict will be created and returned. Default: None. - prefix ([str](https://docs.python.org/3/builtins/stdtypes.html#str), optional) – a prefix added to parameter and buffer names to compose the keys in state_dict. Default: ''. - keep_vars ([bool](https://docs.python.org/3/builtins/functions.html#bool), optional) – by default the [Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor) s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False. Returns: a dictionary containing a whole state of the module Return type: [dict](https://docs.python.org/3/builtins/stdtypes.html#dict) Example: ``` >>> # xdoctest: +SKIP("undefined vars") >>> module.state_dict().keys() ['bias', 'weight'] ``` ```python to(*args, **kwargs) ``` Move and/or cast the parameters and buffers. This can be called as ```python to(device=None, dtype=None, non_blocking=False) ``` ```python to(dtype, non_blocking=False) ``` ```python to(tensor, non_blocking=False) ``` ```python to(memory_format=tensorplay.channels_last) ``` Its signature is similar to tensorplay.Tensor.to(), but only accepts floating point or complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_blocking is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices. See below for examples. > **Note** > > This method modifies the module in-place. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – the desired device of the parameters and buffers in this module - dtype ([tensorplay.dtype](/docs/generated/tensorplay.DType.html#tensorplay.DType)) – the desired floating point or complex dtype of the parameters and buffers in this module - tensor ([tensorplay.Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module - memory_format ([tensorplay.memory_format](/docs/generated/tensorplay.MemoryFormat.html#tensorplay.MemoryFormat)) – the desired memory format for 4D parameters and buffers in this module (keyword only argument) Returns: self Return type: Module Examples: ``` >>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(tensorplay.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=tensorplay.float64) >>> # xdoctest: +REQUIRES(env:TENSORPLAY_DOCTEST_CUDA1) >>> gpu1 = tensorplay.device("cuda:1") >>> linear.to(gpu1, dtype=tensorplay.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16, device='cuda:1') >>> cpu = tensorplay.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=tensorplay.float16) >>> linear = nn.Linear(2, 2, bias=None).to(tensorplay.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=tensorplay.complex128) >>> linear(tensorplay.ones(3, 2, dtype=tensorplay.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=tensorplay.complex128) ``` ```python to_empty(*, device: str | device | int | None, recurse: bool = True) → Self ``` Move the parameters and buffers to the specified device without copying storage. Parameters: - device ([tensorplay.device](/docs/generated/tensorplay.Device.html#tensorplay.Device)) – The desired device of the parameters and buffers in this module. - recurse ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – Whether parameters and buffers of submodules should be recursively moved to the specified device. Returns: self Return type: Module ```python train(mode: bool = True) → Self ``` Set the module in training mode. This has an effect only on certain modules. See the documentation of particular modules for details of their behaviors in training/evaluation mode, i.e., whether they are affected, e.g. Dropout, BatchNorm, etc. Parameters: mode ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – whether to set training mode (True) or evaluation mode (False). Default: True. Returns: self Return type: Module ```python type(dst_type: dtype | str) → Self ``` Casts all parameters and buffers to dst_type. > **Note** > > This method modifies the module in-place. Parameters: dst_type ([type](https://docs.python.org/3/builtins/functions.html#type) or string) – the desired type Returns: self Return type: Module ```python zero_grad(set_to_none: bool = True) → None ``` Reset gradients of all model parameters. See similar function under tensorplay.optim.Optimizer for more context. Parameters: set_to_none ([bool](https://docs.python.org/3/builtins/functions.html#bool)) – instead of setting to zero, set the grads to None. See tensorplay.optim.Optimizer.zero_grad() for details. ## Variables 7 [#](#api-tensorplay.ao.quantization.default_dynamic_qconfig) ### default_dynamic_qconfig data[Full reference ↗](/docs/generated/tensorplay.ao.quantization.default_dynamic_qconfig.html) ```python tensorplay.ao.quantization.default_dynamic_qconfig = (PlaceholderObserver(dtype=tensorplay.float32), MinMaxObserver(dtype=tensorplay.int8, quant_min=-128, quant_max=127)) ``` observer factories for activations and weights. Both fields must be observer classes or callables returning observer instances. Use MyObserver.with_args(x=1) to override constructor arguments. Type: Describes how to quantize a layer [#](#api-tensorplay.ao.quantization.default_dynamic_quant_observer) ### default_dynamic_quant_observer data[Full reference ↗](/docs/generated/tensorplay.ao.quantization.default_dynamic_quant_observer.html) ```python tensorplay.ao.quantization.default_dynamic_quant_observer = PlaceholderObserver(dtype=tensorplay.float32) ``` observer classes be specialized with constructor arguments while staying callable as observer_cls(**kwargs). [#](#api-tensorplay.ao.quantization.default_observer) ### default_observer data[Full reference ↗](/docs/generated/tensorplay.ao.quantization.default_observer.html) ```python tensorplay.ao.quantization.default_observer = MinMaxObserver(quant_min=0, quant_max=127) ``` observer classes be specialized with constructor arguments while staying callable as observer_cls(**kwargs). [#](#api-tensorplay.ao.quantization.default_per_channel_qconfig) ### default_per_channel_qconfig data[Full reference ↗](/docs/generated/tensorplay.ao.quantization.default_per_channel_qconfig.html) ```python tensorplay.ao.quantization.default_per_channel_qconfig = (MovingAverageMinMaxObserver(dtype=tensorplay.qint8), PerChannelMinMaxObserver(ch_axis=0)) ``` observer factories for activations and weights. Both fields must be observer classes or callables returning observer instances. Use MyObserver.with_args(x=1) to override constructor arguments. Type: Describes how to quantize a layer [#](#api-tensorplay.ao.quantization.default_qconfig) ### default_qconfig data[Full reference ↗](/docs/generated/tensorplay.ao.quantization.default_qconfig.html) ```python tensorplay.ao.quantization.default_qconfig = (MinMaxObserver(quant_min=0, quant_max=127), MinMaxObserver(dtype=tensorplay.int8, quant_min=-128, quant_max=127)) ``` observer factories for activations and weights. Both fields must be observer classes or callables returning observer instances. Use MyObserver.with_args(x=1) to override constructor arguments. Type: Describes how to quantize a layer [#](#api-tensorplay.ao.quantization.default_weight_observer) ### default_weight_observer data[Full reference ↗](/docs/generated/tensorplay.ao.quantization.default_weight_observer.html) ```python tensorplay.ao.quantization.default_weight_observer = MinMaxObserver(dtype=tensorplay.int8, quant_min=-128, quant_max=127) ``` observer classes be specialized with constructor arguments while staying callable as observer_cls(**kwargs). [#](#api-tensorplay.ao.quantization.default_weight_only_qconfig) ### default_weight_only_qconfig data[Full reference ↗](/docs/generated/tensorplay.ao.quantization.default_weight_only_qconfig.html) ```python tensorplay.ao.quantization.default_weight_only_qconfig = (PlaceholderObserver(dtype=tensorplay.qint8, is_dynamic=False), PerChannelMinMaxObserver(ch_axis=0)) ``` observer factories for activations and weights. Both fields must be observer classes or callables returning observer instances. Use MyObserver.with_args(x=1) to override constructor arguments. Type: Describes how to quantize a layer