🐛 Describe the bug It came as a surprise to me that setting PYTORCH_NO_CUDA_MEMORY_CACHING=0 is treated the same as PYTORCH_NO_CUDA_MEMORY_CACHING=1 in the native allocator code. The python code of torch handles this envvar by checking i...
Sign in Sign up pytorch / pytorch Public Notifications Fork 22.2k Star 82.5k Code Issues 5k+ Pull requests 1k Actions Projects 12 Wiki Security 1 Insights BC Lint raw_alloc ignores PYTORCH_NO_CUDA_MEMORY_CACHING #170662 Sign in to view logs Summary Jobs bc_linter Run details ...
Tensors and Dynamic neural networks in Python with strong GPU acceleration - Let PYTORCH_NO_CUDA_MEMORY_CACHING has effect only when value is 1 (#… · pytorch/pytorch@8a6e9a8
static const char* cuda_env = getenv("PYTORCH_NO_CUDA_MEMORY_CACHING"); static const char* rocm_env = getenv("PYTORCH_NO_HIP_MEMORY_CACHING"); static bool force_uncached = (cuda_env != nullptr) || (rocm_env != nullptr); bool forceUncachedAllocator() { ...
Tensors and Dynamic neural networks in Python with strong GPU acceleration - Let PYTORCH_NO_CUDA_MEMORY_CACHING has effect only when value is 1 · pytorch/pytorch@887fedc
pytorchmergebot labeled #131114 Status Success Total duration 56s Artifacts – check-labels.yml on: pull_request_target Check labels 43s Oh hello! Nice to see you. Made with ️ by humans.txt Annotations 1 warning Check labels The following actions use a deprecated Node.js...
Tensors and Dynamic neural networks in Python with strong GPU acceleration - raw_alloc ignores PYTORCH_NO_CUDA_MEMORY_CACHING · pytorch/pytorch@19edd4b
Tensors and Dynamic neural networks in Python with strong GPU acceleration - raw_alloc ignores PYTORCH_NO_CUDA_MEMORY_CACHING · pytorch/pytorch@7f9c064
Tensors and Dynamic neural networks in Python with strong GPU acceleration - raw_alloc ignores PYTORCH_NO_CUDA_MEMORY_CACHING · pytorch/pytorch@1a86d8a
Tensors and Dynamic neural networks in Python with strong GPU acceleration - raw_alloc ignores PYTORCH_NO_CUDA_MEMORY_CACHING · pytorch/pytorch@9656a60