ViperEkura 75411ce0cc fix: skip CUDA rotary kernel when grad is enabled ...

- apply_rotary_emb now checks torch.is_grad_enabled() before dispatching to CUDA kernel
- Training (grad enabled) uses torch complex multiply path which supports autograd backward
- Inference (inference_mode/no_grad) uses CUDA kernel as before
- Without this fix, training backward would crash — the CUDA kernel has no autograd backward()

2026-07-31 15:43:15 +08:00

..

lib

refactor: move compiled kernel .so files into extension/lib

2026-07-31 15:36:32 +08:00

__init__.py

perf: add fused CUDA rotary embedding kernel

2026-07-31 15:27:31 +08:00

attention_backend.py

perf: reduce decode overhead in scheduler and executor

2026-07-31 14:50:16 +08:00

attention_ops.py

refactor: unify extension API to blhd layout and is_causal

2026-07-30 18:39:20 +08:00

loader.py

refactor: move compiled kernel .so files into extension/lib

2026-07-31 15:36:32 +08:00

rotary_backend.py

fix: skip CUDA rotary kernel when grad is enabled

2026-07-31 15:43:15 +08:00

rotary_ops.py

perf: add fused CUDA rotary embedding kernel

2026-07-31 15:27:31 +08:00