3 people 616a3e20df [sgl-kernel] Support moe_sum_reduce cuda kernel (#10321) ...

Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
Co-authored-by: Xiaoyu Zhang <35585791+BBuf@users.noreply.github.com>

2025-09-19 14:12:09 +08:00

..

testing

…

__init__.py

[sgl-kernel] Support moe_sum_reduce cuda kernel (#10321)

2025-09-19 14:12:09 +08:00

_fa4_interface.py

support using fa4 on deepseek on blackwell (#9928)

2025-09-16 16:16:06 -07:00

allreduce.py

…

attention.py

…

cutlass_moe.py

…

elementwise.py

[1/2] Speed up trtllm_mla attention backend (>10% e2e) (#10473)

2025-09-15 11:53:21 -07:00

flash_attn.py

Revert "Fix FA4 import cause moe_fused_gate output be illegal memory" (#10432)

2025-09-14 19:03:27 -07:00

fused_moe.py

…

gemm.py

…

grammar.py

…

kvcacheio.py

…

mamba.py

…

marlin.py

…

memory.py

…

moe.py

[sgl-kernel] Support moe_sum_reduce cuda kernel (#10321)

2025-09-19 14:12:09 +08:00

sampling.py

…

scalar_type.py

…

sparse_flash_attn.py

…

spatial.py

…

speculative.py

[Feature] Speculative decoding support lookahead (#9873)

2025-09-18 16:42:41 -07:00

top_k.py

…

utils.py

…

version.py

chore: bump sgl-kernel 0.3.11 (#10630)

2025-09-18 18:43:20 -07:00