Wei-Ming Chen 5b02483c5f AWQ support for vllm fake quant (dense model) (#409) ...

Signed-off-by: weimingc <17592131+meenchen@users.noreply.github.com>

2025-10-09 12:00:49 -07:00

..

backends

Improve realquant gemm impl (#368)

2025-09-26 09:37:51 +05:30

calib

[1/N] QATTrainer training workflow fixes and clean up; Added backend specific unitests; (#318)

2025-09-17 01:43:44 +00:00

nn

TensorQuantizer: remove amax validate forward to reduce cpu-gpu sync; new SVDQuantTensorQuantizer, minor code clean up (#357)

2025-09-26 13:48:10 -07:00

plugins

AWQ support for vllm fake quant (dense model) (#409)

2025-10-09 12:00:49 -07:00

qtensor

Code quality improvements - typos, formatting, etc.

2025-09-02 19:58:33 +05:30

src

Fixed FP8 fake quantization to use fp32 amax scaling; Added support for FP8 per-channel quantization (#381)

2025-09-30 01:56:11 +00:00

triton

Push latest changes and bug fixes (#274)

2025-08-29 17:29:52 +05:30

__init__.py

Add files for 0.27.0 release

2025-04-03 10:31:41 +05:30

algorithms.py

Bug fix for AutoQuantize: Add quantization skipping back; Add disable_layers to the search; cleaner names for search configs (#332)

2025-09-18 15:57:06 -07:00

compress.py

Push latest changes

2025-09-04 11:02:58 +05:30

config.py

Support INT8 Weight-Only Quantization (#263)

2025-09-26 17:46:30 +00:00

conversion.py

[1/N] QATTrainer training workflow fixes and clean up; Added backend specific unitests; (#318)

2025-09-17 01:43:44 +00:00

export_onnx.py

Fix torch.onnx._internal imports for torch 2.9 (#356)

2025-09-23 18:47:42 +00:00

extensions.py

Fixed FP8 fake quantization to use fp32 amax scaling; Added support for FP8 per-channel quantization (#381)

2025-09-30 01:56:11 +00:00

mode.py

Code quality improvements - typos, formatting, etc.

2025-09-02 19:58:33 +05:30

model_calib.py

TensorQuantizer: remove amax validate forward to reduce cpu-gpu sync; new SVDQuantTensorQuantizer, minor code clean up (#357)

2025-09-26 13:48:10 -07:00

model_quant.py

TensorQuantizer: remove amax validate forward to reduce cpu-gpu sync; new SVDQuantTensorQuantizer, minor code clean up (#357)

2025-09-26 13:48:10 -07:00

tensor_quant.py

Fixed FP8 fake quantization to use fp32 amax scaling; Added support for FP8 per-channel quantization (#381)

2025-09-30 01:56:11 +00:00

utils.py

[1/N] QATTrainer training workflow fixes and clean up; Added backend specific unitests; (#318)

2025-09-17 01:43:44 +00:00