Pengchao Wang 8caffd92df [Bugfix][MXFP4] Call trtllm_fp4_block_scale_moe with kwargs (#33104) ...

Signed-off-by: Pengchao Wang <wpc@fb.com>

2026-01-26 15:13:18 -08:00

..

compressed_tensors

[FIX] Always support TP > 4 for FP4 Gemm (#31099)

2026-01-26 11:04:20 -07:00

kernels

[Refactor] Make FP8 Linear Ops use kernel abstraction (#27814)

2026-01-20 14:48:20 +08:00

quark

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

utils

[FIX] Always support TP > 4 for FP4 Gemm (#31099)

2026-01-26 11:04:20 -07:00

__init__.py

[Quantization][Deprecation] Remove RTN (#32697)

2026-01-21 16:34:42 +00:00

awq_marlin.py

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

awq_triton.py

…

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

…

bitblas.py

…

bitsandbytes.py

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

cpu_wna16.py

…

experts_int8.py

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

fbgemm_fp8.py

[Refactor] Make FP8 Linear Ops use kernel abstraction (#27814)

2026-01-20 14:48:20 +08:00

fp8.py

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

fp_quant.py

…

gguf.py

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

gptq_bitblas.py

…

gptq_marlin_24.py

…

gptq_marlin.py

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

gptq.py

…

inc.py

…

input_quant_fp8.py

[torch.compile] Compile CustomOp.forward_native for SiluAndMul and QuantFP8 to avoid raw torch ops inside opaque custom ops (#32806)

2026-01-22 19:52:26 -08:00

ipex_quant.py

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

kv_cache.py

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

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

modelopt.py

[FIX] Always support TP > 4 for FP4 Gemm (#31099)

2026-01-26 11:04:20 -07:00

moe_wna16.py

[MoE Refactor] Move select_experts from FusedMoEQuantMethod -> FusedMoE (#31996)

2026-01-22 18:21:35 -05:00

mxfp4.py

[Bugfix][MXFP4] Call trtllm_fp4_block_scale_moe with kwargs (#33104)

2026-01-26 15:13:18 -08:00

petit.py

…

ptpc_fp8.py

[Refactor] Make FP8 Linear Ops use kernel abstraction (#27814)

2026-01-20 14:48:20 +08:00

qutlass_utils.py

…

schema.py

…

torchao.py

…