amirkl94andmgoin 207b750e19 [NVIDIA] Add SM100 Flashinfer MoE per tensor scale fp8 backend (#21458) ...

Signed-off-by: Amir Klein <203507526+amirkl94@users.noreply.github.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>

2025-07-31 06:00:01 -07:00

..

configs

[Misc] Add fused_moe configs for Qwen3-Coder-480B-A35B-Instruct-FP8 (#21700)

2025-07-27 20:12:18 -07:00

__init__.py

[Kernels] MoE refactor (#19636)

2025-07-02 06:08:27 -07:00

batched_deep_gemm_moe.py

[Kernel][Performance] Tweak MoE Batched silu_mul_fp8_quant_deep_gemm kernel (#21193)

2025-07-18 23:09:51 -07:00

batched_triton_or_deep_gemm_moe.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00

config.py

[Misc] Fix duplicate FusedMoEConfig debug messages (#21455)

2025-07-24 01:27:30 -07:00

cpu_fused_moe.py

[Misc] Add SPDX-FileCopyrightText (#20428)

2025-07-04 07:40:42 +00:00

cutlass_moe.py

Revert "[Performance] Performance improvements in non-blockwise fp8 CUTLASS MoE (#20762) (#21334)

2025-07-21 21:49:01 -07:00

deep_gemm_moe.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00

deep_gemm_utils.py

[Bugfix] DeepGemm utils : Fix hardcoded type-cast (#21517)

2025-07-24 20:17:29 -07:00

deepep_ht_prepare_finalize.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00

deepep_ll_prepare_finalize.py

[Misc] Expand SUPPORTED_HIDDEN_SIZES for DeepEP low-latency kernels (#21818)

2025-07-30 20:41:12 -07:00

flashinfer_cutlass_moe.py

Update fp4 quantize API (#21327)

2025-07-21 23:40:21 -07:00

flashinfer_cutlass_prepare_finalize.py

Update flashinfer CUTLASS MoE Kernel (#21408)

2025-07-24 08:13:31 -07:00

fused_batched_moe.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00

fused_marlin_moe.py

Support Llama 4 for fused_marlin_moe (#20457)

2025-07-04 07:55:10 +00:00

fused_moe.py

[NVIDIA] Add SM100 Flashinfer MoE per tensor scale fp8 backend (#21458)

2025-07-31 06:00:01 -07:00

layer.py

[NVIDIA] Fix Llama4 Scout FP4 functionality issues (#21499)

2025-07-30 07:33:40 -07:00

modular_kernel.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00

moe_align_block_size.py

[Refactor] Remove moe_align_block_size_triton (#21335)

2025-07-26 07:08:29 -07:00

moe_pallas.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

moe_permute_unpermute.py

Fix CUDA permute/unpermute for use with DeepGemm Moe (#17934)

2025-07-27 07:08:00 -07:00

moe_torch_iterative.py

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

pplx_prepare_finalize.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00

prepare_finalize.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00

rocm_aiter_fused_moe.py

[V1] [ROCm] Enable EP with AITER Fused MoE (#20270)

2025-07-01 16:48:30 +00:00

topk_weight_and_reduce.py

[Misc] ModularKernel : Perform WeightAndReduce inside TritonExperts & DeepGemmExperts (#20725)

2025-07-14 19:47:16 +00:00

triton_deep_gemm_moe.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00

utils.py

[Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037)

2025-07-17 21:32:45 -07:00