Cheng Wan 9c138a0445 [3/N] MoE Refactor: Simplify DeepEP Output (#8421)

2025-07-28 11:37:17 -07:00

..

ep_moe

[3/N] MoE Refactor: Simplify DeepEP Output (#8421)

2025-07-28 11:37:17 -07:00

fused_moe_triton

Support triton kernels v3.4.0 for fused_moe (#8258)

2025-07-27 02:28:49 -07:00

token_dispatcher

[3/N] MoE Refactor: Simplify DeepEP Output (#8421)

2025-07-28 11:37:17 -07:00

cutlass_moe_params.py

[CUTLASS-FP4-MOE] Introduce CutlassMoEParams class for easy initialization of Cutlass Grouped Gems Metadata (#6887)

2025-06-05 13:13:14 -07:00

cutlass_moe.py

Add a CUDA kernel for fusing mapping and weighted sum for MoE. (#6916)

2025-06-07 15:24:39 -07:00

cutlass_w4a8_moe.py

feat: support DeepSeek-R1-W4AFP8 model with ep-moe mode (#7762)

2025-07-07 14:47:21 -07:00

fused_moe_native.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00

router.py

Improve streaming, log_level, memory report, weight loading, and benchmark script (#7632)

2025-06-29 23:16:19 -07:00

topk.py

fix:reorder topk experts to ensure shared expert replaces minimal score (#8125)

2025-07-28 20:36:46 +08:00