Roger Wang 4fb8beefaa [Bugfix] Fix cutlass fp8 kernel on hopper for Qwen3.5 (#34914) ...

Signed-off-by: Roger Wang <hey@rogerw.io>

2026-02-19 13:34:55 -08:00

..

configs

…

__init__.py

…

allspark_utils.py

…

flashinfer_fp4_moe.py

[Bugfix] Fix NVFP4 TRTLLM MoE non-gated support; add gsm8k for Nemotron-3-Nano FP8+NVFP4 (#34725)

2026-02-18 09:39:22 -08:00

flashinfer_mxint4_moe.py

…

flashinfer_utils.py

[Bugfix] Fix cutlass fp8 kernel on hopper for Qwen3.5 (#34914)

2026-02-19 13:34:55 -08:00

fp8_utils.py

[Bugfix] Fix quant RMS norm fusion for quantization with TMA-aligned scales (#33255)

2026-02-17 23:35:04 -08:00

gptq_utils.py

…

int8_utils.py

…

layer_utils.py

…

machete_utils.py

…

marlin_utils_fp4.py

…

marlin_utils_fp8.py

…

marlin_utils_test.py

…

marlin_utils.py

…

mxfp4_utils.py

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

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

mxfp6_utils.py

…

mxfp8_utils.py

Add support for ModelOpt MXFP8 dense models (#33786)

2026-02-08 11:16:48 -08:00

nvfp4_emulation_utils.py

…

nvfp4_utils.py

…

ocp_mx_utils.py

[ROCm][Quantization] GPT_OSS in amd-quark format model loading and emulations (#29008)

2026-02-10 10:08:05 -05:00

petit_utils.py

…

quant_utils.py

…

w8a8_utils.py

…