Matthew Bonanni 19fe1a0510

[Kernel] Add FP8 support with FlashMLA backend (#22668) ...

Signed-off-by: Matthew Bonanni <mbonanni001@gmail.com>

2025-08-22 02:26:32 +00:00

..

__init__.py

[Core] Refactor Attention Take 2 (#3462)

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

chunked_prefill_paged_decode.py

Remove unneeded ROCm platform import when using CUDA (#22765)

2025-08-12 21:26:38 -07:00

flashmla.py

[Kernel] Add FP8 support with FlashMLA backend (#22668)

2025-08-22 02:26:32 +00:00

merge_attn_states.py

[Misc] Add SPDX-FileCopyrightText (#19100)

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

nki_flash_attn.py

[Refactor] Remove duplicate ceil_div (#20023)

2025-06-25 05:19:09 +00:00

paged_attn.py

[Misc] Add SPDX-FileCopyrightText (#19100)

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

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

[Bugfix] Make condition in triton kernel constexpr (#22370)

2025-08-06 10:00:58 -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

[Misc] Add SPDX-FileCopyrightText (#19100)

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

triton_decode_attention.py

[Bugfix] Add proper comparison for package versions (#22314)

2025-08-06 20:31:03 -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_unified_attention.py

[BugFix] Fix triton compile error in kernel_unified_attention_2/3d caused by attention sinks (#22368)

2025-08-06 09:47:38 -07:00