jZddlZddlmZmZgdZgdZdZdZejdZ dS) N) check_errorcudart)startstopprofile)gpustarttimestampgpuendtimestamp gridsize3dthreadblocksizestreamidzenableonstart 0conckerneltracec`ttdS)zStarts cuda profiler data collection. .. warning:: Raises CudaError in case of it is unable to start the profiler. N)rrcudaProfilerStart]/lsinfo/ai/hellotax_ai/base_platform/venv/lib/python3.11/site-packages/torch/cuda/profiler.pyrrs( **,,-----rc`ttdS)zStops cuda profiler data collection. .. warning:: Raises CudaError in case of it is unable to stop the profiler. N)rrcudaProfilerStoprrrrrs( ))++,,,,,rc#rK tdVtdS#twxYw)a Enable profiling. Context Manager to enabling profile collection by the active profiling tool from CUDA backend. Example: >>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_CUDA) >>> import torch >>> model = torch.nn.Linear(20, 30).cuda() >>> inputs = torch.randn(128, 20).cuda() >>> with torch.cuda.profiler.profile() as prof: ... model(inputs) N)rrrrrrr&s8   s&6) contextlibrr__all__ DEFAULT_FLAGSrrcontextmanagerrrrrrs!!!!!!!! ' & & ...--- r