yugong333andGitHub ffe1fc7a28 Reduce the kernel overhead when num of active loras is smaller than max loras. Multiple cuda graphs are captured for each num of active-loras. (#32005) ...

Signed-off-by: Yu Gong <yu3.gong@gmail.com>

2026-02-02 12:30:06 -05:00

..

__init__.py

[Core] Allow full cudagraph with separate attention routines and orthogonal to compilation, add support for FA2 and FlashInfer (#20059)

2025-08-15 10:01:39 -04:00

test_cudagraph_dispatch.py

Reduce the kernel overhead when num of active loras is smaller than max loras. Multiple cuda graphs are captured for each num of active-loras. (#32005)

2026-02-02 12:30:06 -05:00

test_cudagraph_mode.py

[Attention] Update tests to remove deprecated env vars (#30563)

2025-12-17 09:49:59 -08:00