lizexu123 5f612a348d [BugFix] fix flashinfer-cutedsl moe nvfp4 (#7120) ...

* fix nvfp4

* fix

* add document

* fix nvfp4

* support eb5

* support bka

* support eb5

* support xpu

* fix

* fix

* add import cutedsl

* fix

* fix

* fix test

* fix H卡

* update document

* fix

* update document

* update document

* fix

2026-04-03 15:43:19 +08:00

..

__init__.py

support w4afp8 EP inference (#3044)

2025-08-25 11:27:45 +08:00

ep.py

[Feature] Support NVFP4 Flashinfer-cutedsl MoE on SM100 (#6963)

2026-03-30 11:37:04 +08:00

flashinfer_cutedsl_moe.py

[BugFix] fix flashinfer-cutedsl moe nvfp4 (#7120)

2026-04-03 15:43:19 +08:00

fused_moe_backend_base.py

[Feature] support compute shared experts before combine for better overlap (#6697)

2026-03-17 15:18:51 +08:00

fused_moe_cutlass_backend.py

[Iluvatar] refactor attn and moe code (#6887)

2026-03-18 10:31:00 +08:00

fused_moe_deepgemm_backend.py

fix bug in cudagraph (#7069)

2026-03-30 14:24:23 +08:00

fused_moe_marlin_backend.py

[Feature] support compute shared experts before combine for better overlap (#6697)

2026-03-17 15:18:51 +08:00

fused_moe_triton_backend.py

[RL] Support moe_topk_select using Paddle native operators and Add fused stack-transpose-quant for BlockWiseFP8 MoE weight quantization and swiglu-fp8-quant op for DeepGemmFusedMoE for training alignment (#6850)

2026-03-24 11:12:39 +08:00

fused_moe_wint2_backend.py

[Feature] support compute shared experts before combine for better overlap (#6697)

2026-03-17 15:18:51 +08:00

moe.py

remove load_up_proj_weight_first (#6932)

2026-03-19 17:21:34 +08:00

routing_indices_cache.py

[RL] add stream guard (#6814)

2026-03-13 11:22:26 +08:00

triton_moe_kernels.py

[Optimization] Use a separate driver when using Triton with Paddle (#6897)

2026-03-24 10:56:00 +08:00