From c701eaff7d69c46a57172b0547bfe2fc05164a0c Mon Sep 17 00:00:00 2001 From: Kaihui-intel <kaihui.tang@intel.com> Date: Fri, 8 Mar 2024 11:51:51 +0800 Subject: [PATCH] fix 3x gptq acc issue (#1654) Signed-off-by: Tang, Kaihui <kaihui.tang@intel.com> --- .../language-modeling/quantization/llm/run_clm_no_trainer.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/llm/run_clm_no_trainer.py b/examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/llm/run_clm_no_trainer.py index 06fdfe0cacf..581db68ddae 100644 --- a/examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/llm/run_clm_no_trainer.py +++ b/examples/3.x_api/pytorch/nlp/huggingface_models/language-modeling/quantization/llm/run_clm_no_trainer.py @@ -328,7 +328,7 @@ def run_fn_for_gptq(model, dataloader_for_calibration, *args): ) quant_config.set_local("lm_head", GPTQConfig(dtype="fp32")) user_model = quantize( - model=user_model, quant_config=quant_config, run_fn=run_fn_for_gptq, run_args=dataloader_for_calibration + model=user_model, quant_config=quant_config, run_fn=run_fn_for_gptq, run_args=(dataloader_for_calibration, ) ) else: # TODO: smooth quant