export CUDA_VISIBLE_DEVICES=0
export VLLM_WORKER_MULTIPROC_METHOD=spawn

MODEL=INCModel2/Qwen3.6-27B-MXFP4-HMT-AutoRound
MODEL_ARGS="pretrained=$MODEL,dtype=bfloat16,enable_thinking=False"

lm_eval \
  --model hf \
  --model_args $MODEL_ARGS \
  --tasks gsm8k \
  --batch_size 64 \
  --gen_kwargs "max_gen_toks=2048" \
  --apply_chat_template \
  --fewshot_as_multiturn \
  --output_path lm_eval_results

Generate the Model

bash

from auto_round import AutoRound

# Load a model (supports FP8/BF16/FP16/FP32)
model_name_or_path = "Qwen/Qwen3.6-27B"
output_dir = "./Qwen3.6-27B-MXFP4-HMT-AutoRound"

ar = AutoRound(model_name_or_path, scheme="MXFP4", iters=200, rotation_config="hadamard")

ar.quantize_and_save(output_dir=output_dir, format="auto_round")

Ethical Considerations and Limitations

The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs.

Therefore, before deploying any applications of the model, developers should perform safety testing.

Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.

Here are a couple of useful links to learn more about Intel's AI software:

Intel Neural Compressor

Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.

Cite

@article{cheng2023optimize, title={Optimize weight rounding via signed gradient descent for the quantization of llms}, author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi}, journal={arXiv preprint arXiv:2309.05516}, year={2023} }

arxiv github