Qwen3.5-4B-Base-AutoRound-W4A16-RTN

Model Details

This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of Qwen/Qwen3.5-4B-Base generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	Qwen/Qwen3.5-4B-Base
Quantization Tool	AutoRound
Quantization Scheme	W4A16
Quantized Size	3697 MB

Evaluation Results

Task	Accuracy
hellaswag	0.5544
mmlu	0.7270
mmlu_abstract_algebra	0.5300
mmlu_anatomy	0.7259
mmlu_astronomy	0.8816
mmlu_business_ethics	0.7500
mmlu_clinical_knowledge	0.8000
mmlu_college_biology	0.8681
mmlu_college_chemistry	0.5600
mmlu_college_computer_science	0.7300
mmlu_college_mathematics	0.4800
mmlu_college_medicine	0.7514
mmlu_college_physics	0.5980
mmlu_computer_security	0.8000
mmlu_conceptual_physics	0.8213
mmlu_econometrics	0.6404
mmlu_electrical_engineering	0.8207
mmlu_elementary_mathematics	0.7645
mmlu_formal_logic	0.5635
mmlu_global_facts	0.4200
mmlu_high_school_biology	0.9000
mmlu_high_school_chemistry	0.7635
mmlu_high_school_computer_science	0.8400
mmlu_high_school_european_history	0.8182
mmlu_high_school_geography	0.8939
mmlu_high_school_government_and_politics	0.9119
mmlu_high_school_macroeconomics	0.7718
mmlu_high_school_mathematics	0.5259
mmlu_high_school_microeconomics	0.8992
mmlu_high_school_physics	0.6821
mmlu_high_school_psychology	0.9028
mmlu_high_school_statistics	0.7269
mmlu_high_school_us_history	0.8578
mmlu_high_school_world_history	0.8565
mmlu_human_aging	0.7085
mmlu_human_sexuality	0.8092
mmlu_humanities	0.6298
mmlu_international_law	0.8347
mmlu_jurisprudence	0.8148
mmlu_logical_fallacies	0.7975
mmlu_machine_learning	0.5893
mmlu_management	0.8835
mmlu_marketing	0.9145
mmlu_medical_genetics	0.8900
mmlu_miscellaneous	0.8327
mmlu_moral_disputes	0.7457
mmlu_moral_scenarios	0.4045
mmlu_nutrition	0.8105
mmlu_other	0.7676
mmlu_philosophy	0.7492
mmlu_prehistory	0.7901
mmlu_professional_accounting	0.5887
mmlu_professional_law	0.5287
mmlu_professional_medicine	0.7868
mmlu_professional_psychology	0.7663
mmlu_public_relations	0.7182
mmlu_security_studies	0.7714
mmlu_social_sciences	0.8271
mmlu_sociology	0.8955
mmlu_stem	0.7342
mmlu_us_foreign_policy	0.8900
mmlu_virology	0.5663
mmlu_world_religions	0.8187
piqa	0.7758

How to Use

HF Usage

Step 1: Install AutoRound

pip install auto-round

Step 2: Load and run the quantized model

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen3.5-4B-Base-AutoRound-W4A16-RTN"

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")

# prepare the model input
prompt = "Write a quick sort algorithm."
messages = [{"role": "user", "content": prompt}]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(**model_inputs, max_new_tokens=512)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]) :].tolist()

content = tokenizer.decode(output_ids, skip_special_tokens=True)
print("content:", content)

VLLM Usage

vllm serve Qwen3.5-4B-Base-AutoRound-W4A16-RTN \
    --trust-remote-code \
    --dtype bfloat16 \
    --tensor_parallel_size 1

If you encounter any issues, feel free to open an issue on the AutoRound GitHub repo or provide feedback on the Low-Bit Open LLM Leaderboard.

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
• AutoRound

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

---

This model is part of the Intel Low-Bit Open LLM Leaderboard initiative.