3 people 87b4d1557d [CustomOp][MM] Extract MMEncoderAttention as CustomOp and replace the backend of QwenVisionAttention with it. (#30125) ...

Signed-off-by: shen-shanshan <467638484@qq.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: tjtanaa <tunjian.tan@embeddedllm.com>

2025-12-15 11:13:32 +08:00

..

__init__.py

[Core] Refactor Attention Take 2 (#3462)

2024-03-25 04:39:33 +00:00

chunked_prefill_paged_decode.py

Convert formatting to use ruff instead of yapf + isort (#26247)

2025-10-05 07:06:22 -07:00

common.py

bugfix: correct attn output with base 2 or e (#28840)

2025-11-29 07:52:12 +08:00

flashmla.py

[Bugfix] fixes the decoding metadata of dense mla's fp8 kvcache. (#27144)

2025-10-21 18:27:03 +00:00

merge_attn_states.py

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

2025-10-12 09:51:31 -07:00

paged_attn.py

[V0 deprecation] Clean up legacy paged attention helper functions (#28043)

2025-11-28 16:44:33 +00:00

pallas_kv_cache_update.py

[Chore]:Extract math and argparse utilities to separate modules (#27188)

2025-10-26 04:03:32 -07:00

prefix_prefill.py

[Misc] Remove unused attention prefix prefill ops functions (#26971)

2025-11-11 18:26:04 +00:00

rocm_aiter_mla_sparse.py

[ROCm] Add AMD GPU support on Deepseek v3.2 and SparseMLA (#26670)

2025-11-20 02:54:01 -08:00

triton_decode_attention.py

Convert formatting to use ruff instead of yapf + isort (#26247)

2025-10-05 07:06:22 -07:00

triton_merge_attn_states.py

[Performance][MLA][ROCm] Remove redundant D2D copy in deepseek (#27457)

2025-11-26 12:45:28 +08:00

triton_reshape_and_cache_flash.py

[Bugfix][CI/Test][Spec Decode] Fix illegal memory access in offline_inference/spec_decode.py (Issue 27619) (#28432)

2025-11-13 22:34:01 -08:00

triton_unified_attention.py

[Kernel] Support CUDA Graphs in 3D Triton Attention Kernel (#28306)

2025-12-12 16:55:40 +01:00

vit_attn_wrappers.py

[CustomOp][MM] Extract MMEncoderAttention as CustomOp and replace the backend of QwenVisionAttention with it. (#30125)

2025-12-15 11:13:32 +08:00