youkaichao a2e6fa7e03 [bugfix][deepseek] fix flashmla kernel selection (#25956) ...

Signed-off-by: youkaichao <youkaichao@gmail.com>

2025-10-01 00:30:36 +08:00

..

__init__.py

[Core] Refactor Attention Take 2 (#3462)

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

chunked_prefill_paged_decode.py

[torch.compile][ROCm][V1] Enable attention output FP8 fusion for V1 attention backends (#19767)

2025-09-10 13:59:55 -07:00

common.py

[New Model] DeepSeek-V3.2 (Rebased to Main) (#25896)

2025-09-30 17:14:41 +08:00

flashmla.py

[bugfix][deepseek] fix flashmla kernel selection (#25956)

2025-10-01 00:30:36 +08:00

merge_attn_states.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

paged_attn.py

[New Model] DeepSeek-V3.2 (Rebased to Main) (#25896)

2025-09-30 17:14:41 +08:00

pallas_kv_cache_update.py

[TPU] kv cache update kernel doesn't need to be padded slices to multiple of num_slices_per_block (#22394)

2025-08-09 20:49:04 -07:00

prefix_prefill.py

[torch.compile][ROCm][V1] Enable attention output FP8 fusion for V1 attention backends (#19767)

2025-09-10 13:59:55 -07:00

rocm_aiter_mla.py

[Kernel] Apply torch.Tag.needs_fixed_stride_order only for torch==2.6.0 (#19346)

2025-07-18 14:10:21 -04:00

rocm_aiter_paged_attn.py

Aiter mha fp8 fix (#24991)

2025-09-17 22:29:14 +00:00

triton_decode_attention.py

[Kernel] Support DCP for Triton backend (#25132)

2025-09-24 18:09:34 -07:00

triton_flash_attention.py

[Misc] Fix misleading ROCm warning (#19486)

2025-06-12 00:12:10 +08:00

triton_merge_attn_states.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

triton_reshape_and_cache_flash.py

[Misc] fix tests failure by using current_platform (#25825)

2025-09-29 04:18:57 +00:00

triton_unified_attention.py

Optimize triton unified attention performance for sliding window attention (#24390)

2025-09-19 13:07:26 -06:00