danisereb 0a6a3a1290 Add support for ModelOpt MXFP8 MoE models (#35986) ...

Signed-off-by: Daniel Serebrenik <daserebrenik@nvidia.com>

2026-03-08 13:00:05 -07:00

..

configs

[Model Performance] Add Qwen3MoE tuned MoE configs for H200 (#35457)

2026-02-27 13:51:14 +08:00

experts

[Bugfix] Fix passing of activation_type to trtllm fused MoE NVFP4 and FP8 (#36017)

2026-03-04 22:23:51 +00:00

oracle

Add support for ModelOpt MXFP8 MoE models (#35986)

2026-03-08 13:00:05 -07:00

prepare_finalize

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

router

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

runner

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

__init__.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

activation.py

[Refactor] Replace activation: str with MoEActivation enum (#33843)

2026-02-11 17:29:32 -08:00

all2all_utils.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

batched_deep_gemm_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

config.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

cpu_fused_moe.py

[Bugfix] fix activation in cpu_fused_moe_torch call (#34696)

2026-02-17 23:39:46 -08:00

cutlass_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

deep_gemm_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

deep_gemm_utils.py

[Perf] Optimize deepgemm experts initialization, 3.9% TTFT improvement (#30494)

2025-12-11 17:28:34 -05:00

deepep_ht_prepare_finalize.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

deepep_ll_prepare_finalize.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

fallback.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

flashinfer_a2a_prepare_finalize.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

flashinfer_cutedsl_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

flashinfer_cutlass_moe.py

[Bugfix][CI] fix typos (#34934)

2026-03-05 17:05:46 +00:00

flashinfer_trtllm_moe.py

[Bugfix] Disable FlashInfer TRTLLM BF16 path for non-gated MoE (#36146)

2026-03-05 09:49:05 -08:00

fused_batched_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

fused_marlin_moe.py

[Bugfix][CI] fix typos (#34934)

2026-03-05 17:05:46 +00:00

fused_moe_method_base.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

fused_moe_modular_method.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

fused_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

gpt_oss_triton_kernels_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

layer.py

Add support for ModelOpt MXFP8 MoE models (#35986)

2026-03-08 13:00:05 -07:00

modular_kernel.py

[Perf] Optimize FusedMoEModularKernel output tensor using torch.empty (#35794)

2026-03-05 13:47:53 -05:00

moe_align_block_size.py

[Kernel][MoE] optimize moe_align_block_size (#29642)

2025-12-07 01:58:47 -08:00

moe_permute_unpermute.py

[Refactor] Remove align block size logic in moe_permute (#33449)

2026-02-06 10:57:06 -08:00

mori_prepare_finalize.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

rocm_aiter_fused_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

routed_experts_capturer.py

[Bugfix] fix device_name for routing replay (#34336)

2026-02-26 12:18:38 +00:00

shared_fused_moe.py

[MoE Refactor] Introduce MoERunner abstraction and move execution logic from FusedMoE to DefaultMoERunner (#32344)

2026-02-10 19:51:07 -05:00

topk_weight_and_reduce.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

triton_cutlass_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

triton_deep_gemm_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

trtllm_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

unquantized_fused_moe_method.py

[Hardware] Replace torch.cuda.empty_cache with torch.accelerator.empty_cache (#30681)

2026-03-04 09:49:47 +00:00

utils.py

[MoE Refactor] MXFP4 Cutlass Experts to MK (#34542)

2026-02-25 17:32:39 -08:00

xpu_fused_moe.py

[MoE Refactor] Create MK for TRTLLM Kernels (#32564)

2026-03-03 10:39:50 -08:00

zero_expert_fused_moe.py

[MoE Refactor][5/N] Isolate zero expert to LongCatFlash (#28891)

2025-12-20 18:22:04 +00:00