Chao-Ju ChenandMichael Goin 245758992e [Bugfix] Rescale NVFP4 weight scales to fix BF16 dequant underflow (#34577) ...

Signed-off-by: ricky-chaoju <ricky.chen@infinirc.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>

2026-03-17 20:48:42 +00:00

..

compressed_tensors

Fix eplb nvfp4 experts hook (#37217)

2026-03-16 22:03:54 +00:00

quark

Fix infinite recursive search issue in quark.py (#32779)

2026-03-17 07:19:15 +00:00

utils

[Bugfix] Rescale NVFP4 weight scales to fix BF16 dequant underflow (#34577)

2026-03-17 20:48:42 +00:00

__init__.py

[Quant][Feature] Support online MXFP8 quantization for MoE and dense models (#35448)

2026-03-16 18:07:39 -04:00

awq_marlin.py

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

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

awq_triton.py

Convert formatting to use ruff instead of yapf + isort (#26247)

2025-10-05 07:06:22 -07:00

awq.py

[Bugfix] Fix missing is_layer_skipped check for FusedMoE in AWQConfig (#32935)

2026-01-23 17:19:56 -05:00

base_config.py

[Quantization] - Added uses_meta_device_weights to quant config (#34645)

2026-02-17 23:43:44 -08:00

bitsandbytes.py

[ROCm] Enable bitsandbytes quantization support on ROCm (#34688)

2026-02-21 00:34:55 -08:00

cpu_wna16.py

[Torch 2.11] Guard torch._C._cpu attribute checks for forward compatibility (#35673)

2026-03-17 18:47:59 +00:00

experts_int8.py

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

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

fbgemm_fp8.py

[Refactor] [1/N] Reorganize kernel abstraction directory (#34055)

2026-02-24 06:47:22 +00:00

fp8.py

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

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

fp_quant.py

"refactor: refactor_repeated_interfaces" (#32486)

2026-01-18 22:07:01 +08:00

gguf.py

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

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

gptq_marlin.py

[Refactor] [1/N] Reorganize kernel abstraction directory (#34055)

2026-02-24 06:47:22 +00:00

gptq.py

[Quantization][Deprecation] Remove BitBlas (#32683)

2026-01-28 11:06:22 +00:00

inc.py

[XPU][1/N] Deprecate ipex and switch to vllm-xpu-kernels for xpu platform (#33379)

2026-02-02 22:46:10 -08:00

input_quant_fp8.py

[Bugfix] Fix Dynamo unexpected keyword argument (#34320)

2026-02-16 01:32:30 -08:00

kv_cache.py

[Bugfix][Attention] Explicitly report support for kv_cache_dtype bfloat16 (#32795)

2026-01-22 19:05:18 +00:00

modelopt.py

[Quant][Feature] Support online MXFP8 quantization for MoE and dense models (#35448)

2026-03-16 18:07:39 -04:00

moe_wna16.py

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

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

mxfp4.py

[2/N] Elastic EP Milestone 2: Integrating NIXL-EP (#35627)

2026-03-13 09:25:33 -04:00

mxfp8.py

[Quant][Feature] Support online MXFP8 quantization for MoE and dense models (#35448)

2026-03-16 18:07:39 -04:00

petit.py

[Misc] Replace Optional[X] with X | None syntax (#33332)

2026-01-30 01:56:59 -08:00

ptpc_fp8.py

[Refactor] Remove dead or duplicate func utils or variables (#35318)

2026-02-26 10:57:56 -05:00

qutlass_utils.py

"refactor: refactor_repeated_interfaces" (#32486)

2026-01-18 22:07:01 +08:00

schema.py

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

2025-10-12 09:51:31 -07:00

torchao.py

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

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