tensorplay.testing.default_tolerances

tensorplay.testing.default_tolerances(*inputs: Tensor | dtype, dtype_precisions: dict[dtype, tuple[float, float]] | None = None) → tuple[float, float][source]

Returns the default absolute and relative testing tolerances for a set of inputs based on the dtype.

Returns:

Loosest tolerances of all input dtypes.

Return type:

(Tuple[float, float])