fxyfxy777 f3413c4caa [BugFix] fix fused_mask_swiglu_fp8_quant bug (#6316) ...

* optimize mask_quant op speed up 1.5

* fix calculate sequence

* add fused

* rm log

* push kernel code

* add ut

* accuracy ok

* add ue8m0

* add ut

* add merge develop

* rm ut of mask_per_token_quant

* Revert "[Optimize] optimize mask_quant & swiglu (#6222)"

This reverts commit 2ada119a38.

* add block_size

* pre-commit

2026-02-03 13:54:12 +08:00

..

__init__.py

support w4afp8 EP inference (#3044)

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

ep.py

Revert "[Feature] Support Ernie FP8 on sm100 (#5593)" (#6275)

2026-01-30 11:22:01 +08:00

fused_moe_backend_base.py

[Feature] Support redundant expert for eplb (#5918)

2026-01-09 17:13:24 +08:00

fused_moe_cutlass_backend.py

[Iluvartar][CI] Fix the error max_tokens_per_expert referenced before assignment (#6083)

2026-01-21 16:01:29 +08:00

fused_moe_deepgemm_backend.py

[BugFix] fix fused_mask_swiglu_fp8_quant bug (#6316)

2026-02-03 13:54:12 +08:00

fused_moe_marlin_backend.py

[New][RL] Support Rollout Routing Replay (#5405)

2025-12-05 22:06:26 +08:00

fused_moe_triton_backend.py

Revert "[Feature] Support Ernie FP8 on sm100 (#5593)" (#6275)

2026-01-30 11:22:01 +08:00

fused_moe_wint2_backend.py

[BugFix] fix wint2 (#6109)

2026-01-20 21:46:21 +08:00

moe.py

[Feature] Support NVFP4 MoE on SM100 (#6003)

2026-01-29 14:16:07 +08:00

routing_indices_cache.py

[UT] Add GLM E2E tests for non-MTP and MTP (#6163)

2026-01-23 10:34:29 +08:00

triton_moe_kernels.py

[OPs] MoE support wfp8afp8(channelwise) and improve per_token_quant_fp8 (#4238)

2025-09-24 16:39:51 +08:00