apex · Issues· 770 open
Open on GitHubLocally synced open issues (discussions stay on GitHub)
- #2024
`torch.cuda.amp` still used in 5 files (incl. private `autocast_mode._cast` / `grad_scaler.OptState`) — deprecated since torch 2.3/2.4, no guards
Updated Aug 22, 2026 - #2005
incompatibility with nightly pytorch as it requires C++20
Updated May 22, 2026 - #833
No speedup on RTX card, how apex affects loss function that uses long float?
Updated May 14, 2026 - #1999
[Bug] FusedRMSNorm leaks 2 CUDA tensors per forward call under torch.no_grad()
bugUpdated Apr 29, 2026 - #1998
CWE-22/CWE-73 in permutation cache path: APEX_ASP_CACHE_DIR controls write destination
Updated Apr 27, 2026 - #1991
cannot import name 'amp' from 'apex'
bugUpdated Apr 27, 2026 - #1686
sequence parallel with rmsnorm/layernorm
Updated Mar 26, 2026 - #368
FileNotFoundError: [Errno 2] No such file or directory: ':/usr/local/cuda:/usr/local/cuda-10.1/bin/nvcc': ':/usr/local/cuda:/usr/local/cuda-10.1/bin/nvcc'
Updated Jan 17, 2026 - #1945
Think about removing `apex_C`
Updated Jan 14, 2026 - #1970
problem size question for multi-tensor-axpby
Updated Dec 15, 2025 - #1964
Reduce the number of ignored rules
Updated Nov 29, 2025 - #1843
ASP Automatic Sparsity forward function For Loop Error
Updated Nov 23, 2025 - #1878
install warning
bugUpdated Nov 23, 2025 - #1946
Migrate to LibTorch Stable ABI
Updated Nov 7, 2025 - #1939
Unable to install apex-ModuleNotFoundError: No module named 'packaging'
Updated Nov 3, 2025