Duncan Moss 552bbe6f4e [Attention] Add head_dim=512 support for FlashInfer trtllm attention backend (#38822)

2026-05-22 20:27:35 -04:00

..

mla

[XPU]fix: add XPU platform guards to DeepSeek-V4 ops (#42950)

2026-05-23 06:29:45 +08:00

__init__.py

[V1] Implement vLLM V1 [1/N] (#9289)

2024-10-22 01:24:07 -07:00

cpu_attn.py

[CPU][RISC-V] Add VLEN=256 support to RVV attention kernels (#42943)

2026-05-21 04:50:49 -07:00

fa_utils.py

[FlashAttn] Fix supports_kv_cache_dtype() accepting unhandled fp8 kv-cache dtype variants (#42685)

2026-05-15 15:34:59 -04:00

flash_attn_diffkv.py

[FlashAttn] Fix supports_kv_cache_dtype() accepting unhandled fp8 kv-cache dtype variants (#42685)

2026-05-15 15:34:59 -04:00

flash_attn.py

[XPU] update xpu graph usage (#43043)

2026-05-19 23:09:07 +08:00

flashinfer.py

[Attention] Add head_dim=512 support for FlashInfer trtllm attention backend (#38822)

2026-05-22 20:27:35 -04:00

flex_attention.py

expose flex block size for batch invariant mode (#41252)

2026-05-13 14:11:57 -07:00

gdn_attn.py

[Models][GDN] Remove GPU/CPU syncs in GDNAttentionMetadata.build during speculative decoding (#38047)

2026-04-06 15:39:37 +00:00

linear_attn.py

[HMA]Fix corner case when hybrid page_size can not be evenly divided issue (blk_size=64,tp=4) (#37467)

2026-03-30 16:47:30 +00:00

mamba1_attn.py

[HMA]Fix corner case when hybrid page_size can not be evenly divided issue (blk_size=64,tp=4) (#37467)

2026-03-30 16:47:30 +00:00

mamba2_attn.py

[Bugfix][Hybrid][NemotronH] Fix mamba_cache_mode=all + speculative decoding crash (#41233)

2026-05-18 14:54:00 +03:00

mamba_attn.py

[Bugfix] mamba: run single-token extends as decodes (#42430)

2026-05-18 15:26:00 +00:00

registry.py

[MLA Attention Backend] Add TOKENSPEED_MLA backend for DSR1/Kimi K25 prefill + decode on Blackwell (#41778)

2026-05-13 23:48:02 -07:00

rocm_aiter_fa.py

[ROCm] Clean up a bit the AITER FA backend (#41942)

2026-05-11 22:45:18 +08:00

rocm_aiter_unified_attn.py

[ROCm] aiter_unified_attn fp8 q scale refactor (#38296)

2026-05-06 16:11:36 +00:00

rocm_attn.py

[ROCm] Cap Triton paged attention block size to fix ROCm shared memory OOM (#38502)

2026-05-10 10:03:00 +00:00

short_conv_attn.py

[HMA]Fix corner case when hybrid page_size can not be evenly divided issue (blk_size=64,tp=4) (#37467)

2026-03-30 16:47:30 +00:00

triton_attn.py

[Bugfix] Source num_qo_heads from Attention layers in Flashinfer/Triton metadata builders (#42650)

2026-05-22 16:10:03 +00:00

turboquant_attn.py

[Perf] zeros -> empty to remove additional fill (#42988)

2026-05-21 16:00:29 -04:00

utils.py

[Bugfix] Source num_qo_heads from Attention layers in Flashinfer/Triton metadata builders (#42650)

2026-05-22 16:10:03 +00:00