nvjullinandMichael Goin f66673a39d [Kernel] Added flashinfer fp8 per-tensor gemms (#22895) ...

Signed-off-by: Julien Lin <jullin@nvidia.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>

2025-08-26 06:54:04 -07:00

..

compressed_tensors

[Perf] Remove duplicated NVFP4 blockscales to save memory (#23379)

2025-08-26 19:16:33 +08:00

kernels

[kernel] Support W4A8 on Hopper (#23198)

2025-08-24 06:18:04 +00:00

quark

[Kernels] Clean up FusedMoeMethodBase and modular kernel setup. Remove extra arguments from modular kernel methods. (#22035)

2025-08-15 14:46:00 -04:00

utils

[Kernel] Added flashinfer fp8 per-tensor gemms (#22895)

2025-08-26 06:54:04 -07:00

__init__.py

Quantization: support FP4 quantized models on AMD CDNA2/CDNA3 GPUs (#22527)

2025-08-22 20:53:21 -06:00

auto_round.py

[Kernels] Clean up FusedMoeMethodBase and modular kernel setup. Remove extra arguments from modular kernel methods. (#22035)

2025-08-15 14:46:00 -04:00

awq_marlin.py

[Kernels] Clean up FusedMoeMethodBase and modular kernel setup. Remove extra arguments from modular kernel methods. (#22035)

2025-08-15 14:46:00 -04:00

awq_triton.py

[Misc] Add SPDX-FileCopyrightText (#19100)

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

awq.py

[Kernels] Clean up FusedMoeMethodBase and modular kernel setup. Remove extra arguments from modular kernel methods. (#22035)

2025-08-15 14:46:00 -04:00

base_config.py

[Quant] [Bugfix] Fix quantization config matching with hf_to_vllm_mapper (#20046)

2025-07-01 19:20:34 +09:00

bitblas.py

[Bugfix] Add proper comparison for package versions (#22314)

2025-08-06 20:31:03 -07:00

bitsandbytes.py

[Kernels] Clean up FusedMoeMethodBase and modular kernel setup. Remove extra arguments from modular kernel methods. (#22035)

2025-08-15 14:46:00 -04:00

deepgemm.py

[Feature] Integrate SM100 DeepGEMM support (#20087)

2025-07-10 20:18:05 -07:00

deepspeedfp.py

[Bugfix] Add proper comparison for package versions (#22314)

2025-08-06 20:31:03 -07:00

experts_int8.py

[Kernels] Clean up FusedMoeMethodBase and modular kernel setup. Remove extra arguments from modular kernel methods. (#22035)

2025-08-15 14:46:00 -04:00

fbgemm_fp8.py

[Perf][fp8] Use CustomOp abstraction for fp8 quant for better perf (#19830)

2025-07-11 04:56:28 +00:00

fp8.py

[Kernel] Added flashinfer fp8 per-tensor gemms (#22895)

2025-08-26 06:54:04 -07:00

gguf.py

[Quantization] Allow GGUF quantization to skip unquantized layer (#23188)

2025-08-22 13:04:22 -06:00

gptq_bitblas.py

[Bugfix] Add proper comparison for package versions (#22314)

2025-08-06 20:31:03 -07:00

gptq_marlin_24.py

[Misc] Add SPDX-FileCopyrightText (#19100)

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

gptq_marlin.py

fix: gptq marlin weight loading failure (#23066)

2025-08-17 15:56:07 -07:00

gptq.py

feat: Add Support GPTQ Quantization MOE on ROCM vllm serve (#21733)

2025-08-01 21:12:19 -04:00

hqq_marlin.py

[Kernel] [Quantization] Add MXFP4 and bias support for marlin kernel (#22428)

2025-08-14 11:23:22 -07:00

inc.py

Support FP8 Quantization and Inference Run on Intel Gaudi (HPU) using INC (Intel Neural Compressor) (#12010)

2025-07-16 15:33:41 -04:00

input_quant_fp8.py

[Perf][fp8] Use CustomOp abstraction for fp8 quant for better perf (#19830)

2025-07-11 04:56:28 +00:00

ipex_quant.py

[Bugfix] Add proper comparison for package versions (#22314)

2025-08-06 20:31:03 -07:00

kv_cache.py

[Misc] Add SPDX-FileCopyrightText (#19100)

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

modelopt.py

[Perf] Remove duplicated NVFP4 blockscales to save memory (#23379)

2025-08-26 19:16:33 +08:00

moe_wna16.py

[Kernels] Clean up FusedMoeMethodBase and modular kernel setup. Remove extra arguments from modular kernel methods. (#22035)

2025-08-15 14:46:00 -04:00

mxfp4.py

Update Flashinfer to 0.2.14.post1 (#23537)

2025-08-25 18:30:44 -07:00

neuron_quant.py

[Misc] Add SPDX-FileCopyrightText (#19100)

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

petit.py

Quantization: support FP4 quantized models on AMD CDNA2/CDNA3 GPUs (#22527)

2025-08-22 20:53:21 -06:00

ptpc_fp8.py

[Kernel] Added flashinfer fp8 per-tensor gemms (#22895)

2025-08-26 06:54:04 -07:00

rtn.py

[Kernels] Clean up FusedMoeMethodBase and modular kernel setup. Remove extra arguments from modular kernel methods. (#22035)

2025-08-15 14:46:00 -04:00

schema.py

[Misc] Add SPDX-FileCopyrightText (#19100)

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

torchao.py

[Quant] [Bugfix] Fix quantization config matching with hf_to_vllm_mapper (#20046)

2025-07-01 19:20:34 +09:00

tpu_int8.py

[TPU] Add support for online w8a8 quantization (#22425)

2025-08-08 23:12:54 -07:00