API symbolscudaCopyCopy as MarkdownCopy page linkView MarkdownDownload .mdOpenAIOpen in ChatGPTClaudeOpen in ClaudePerplexityOpen in PerplexityGrokOpen in Grok tensorplay.cuda.get_allocator_backend tensorplay.cuda.get_allocator_backend() → str[source] Returns the active allocator backend as a string. Always native here. Help improve this pageFound an error, an unclear step, or a missing example?Was this page helpful?YesNoSuggest an editReport an issuetensorplay.cuda.empty_cacheRelease all unoccupied cached memory currently held by the caching allocator so that those can be used in other GPU application and visible in nvidia-smi .tensorplay.cuda.get_arch_listReturn list CUDA architectures this library was compiled for.
