Yuankai Chenandgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> 2e777d21a8 [Bugfix][Rocm]Aiter MoE re-uses existing tensor addresses after weight update. (#40390) ...

Signed-off-by: Yuankai Chen <yuankach@amd.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

2026-05-06 10:32:26 +00:00

..

configs

Add tuned triton fused_moe configs on H100 for gpt-oss (#39904)

2026-04-28 03:38:48 -07:00

experts

[ROCm][Quantization][2/N] Refactor quark_moe w4a8 w/ oracle (#39136)

2026-05-04 19:50:38 -07:00

oracle

[New Model][ROCm] Add AMD support for DeepSeek V4 (#40871)

2026-05-05 08:55:37 -07:00

prepare_finalize

[DSV4] Add BF16 and MXFP8 A2A support for flashinfer a2a one sided (#40960)

2026-04-30 15:33:12 -07:00

router

[DeepSeek] Use torch.mm for bf16xbf16->fp32 gemm (#41300)

2026-04-30 16:28:57 -07:00

runner

[MoE] Make MoERunnerInterface a PluggableLayer for OOT support (#35178)

2026-04-30 03:31:08 -07:00

__init__.py

[MoE] Move cutlass moe to fused_moe/experts/ (#40574)

2026-04-24 06:05:49 +00:00

activation.py

[Model] Fix Gemma4 MoE activation mismatch (#41574)

2026-05-05 04:34:11 +00:00

all2all_utils.py

[DSV4] Add BF16 and MXFP8 A2A support for flashinfer a2a one sided (#40960)

2026-04-30 15:33:12 -07:00

config.py

[DSV4] Add BF16 and MXFP8 A2A support for flashinfer a2a one sided (#40960)

2026-04-30 15:33:12 -07:00

cpu_fused_moe.py

[DSV4] Add silu clamp limit to shared expert (#40950)

2026-04-27 00:37:43 -07:00

deep_gemm_utils.py

Fix DeepGEMM ep_scatter output address overflow (#39213)

2026-05-05 18:56:56 +00:00

fallback.py

[Feature]: Remove Chunking From FusedMoE (#34086)

2026-03-12 14:24:38 -04:00

flashinfer_cutlass_moe.py

[Bugfix] Disable FlashInfer CUTLASS MoE on SM110 (Jetson Thor AGX) (#40808)

2026-04-30 20:08:49 -07:00

fused_batched_moe.py

[Model] Fix Gemma4 MoE activation mismatch (#41574)

2026-05-05 04:34:11 +00:00

fused_humming_moe.py

[Model] Fix Gemma4 MoE activation mismatch (#41574)

2026-05-05 04:34:11 +00:00

fused_marlin_moe.py

[Model] Fix Gemma4 MoE activation mismatch (#41574)

2026-05-05 04:34:11 +00:00

fused_moe_method_base.py

[Feat] DeepSeek V4 Rebased (#40860)

2026-04-26 18:31:08 -07:00

fused_moe_modular_method.py

[MoE Refactor] Make SharedExperts class for use with DefaultMoERunner (#35153)

2026-04-01 09:44:08 -04:00

fused_moe.py

[Model] Fix Gemma4 MoE activation mismatch (#41574)

2026-05-05 04:34:11 +00:00

layer.py

[XPU] enable is_act_and_mul for xpu (#37481)

2026-05-05 01:07:39 +00:00

lora_context.py

[LoRA] MoE LoRA Refactor (#40338)

2026-04-26 01:55:19 +00:00

lora_experts_mixin.py

[LoRA] MoE LoRA Refactor (#40338)

2026-04-26 01:55:19 +00:00

modular_kernel.py

[LoRA] MoE LoRA Refactor (#40338)

2026-04-26 01:55:19 +00:00

moe_align_block_size.py

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

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

moe_fused_mul_sum.py

[Quantization] add humming quantization kernel (#34556)

2026-04-24 09:29:44 -04:00

moe_permute_unpermute.py

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

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

rocm_aiter_fused_moe.py

[ROCm][Bugfix]: W4A4 MOE using emulation instead of AITER on MXFP4-supported hardware (#41175)

2026-04-29 16:39:03 -05:00

routed_experts_capturer.py

[Bugfix] Fix RoutedExpertsCapturer for Gemma 4 MoE (top_k_experts) (#41401)

2026-04-30 16:19:59 -07:00

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] Move cutlass moe to fused_moe/experts/ (#40574)

2026-04-24 06:05:49 +00:00

triton_deep_gemm_moe.py

[MoE] Move DEEP_GEMM into experts/ subdirectory (#39005)

2026-04-08 19:23:08 +00:00

unquantized_fused_moe_method.py

[Bugfix][Rocm]Aiter MoE re-uses existing tensor addresses after weight update. (#40390)

2026-05-06 10:32:26 +00:00

utils.py

[DSV4] Add BF16 and MXFP8 A2A support for flashinfer a2a one sided (#40960)

2026-04-30 15:33:12 -07:00