Jinzhen Lin d74e5f37bc

[Kernel] fp4 marlin kernel (#17687) ...

Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>

2025-05-10 19:58:49 -07:00

..

attention

fix broken test vllm:test_kernels - test_attention_selector.py::test_flash_attn (#17873)

2025-05-10 10:46:54 +08:00

core

Make key optional for rotary embedding (#17566)

2025-05-07 00:11:46 -07:00

mamba

[Model] Mamba2 causal conv1d Refactor to Split Prefill and Decode Requests for Corresponding Kernels (#17146)

2025-05-06 17:59:30 -07:00

moe

[Kernel] fp4 marlin kernel (#17687)

2025-05-10 19:58:49 -07:00

quantization

[Kernel] fp4 marlin kernel (#17687)

2025-05-10 19:58:49 -07:00

__init__.py

…

allclose_default.py

…

quant_utils.py

…

test_cutlass_mla_decode.py

[NVIDIA] Support Cutlass MLA for Blackwell GPUs (#16032)

2025-04-27 06:29:21 -07:00

test_fused_quant_activation.py

[torch.compile] Add torch inductor pass for fusing silu_and_mul with subsequent scaled_fp8_quant operations (#10867)

2025-05-01 07:59:28 -07:00

test_triton_flash_attention.py

[Kernel][Triton][FP8] Adding fp8 and variable length sequence support to Triton FAv2 kernel (#12591)

2025-04-27 00:35:08 +00:00

test_triton_unified_attention.py

[Kernel] Unified Triton kernel that doesn't distinguish between prefill + decode (#16828)

2025-05-06 18:21:48 -04:00

utils.py

…