Qwen3.5-4B-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 generated by AutoRound. Please follow the license of the original model.

Quantization Details

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

Evaluation Results

Task	Accuracy
hellaswag	0.5362
mmlu	0.7272
mmlu_abstract_algebra	0.5300
mmlu_anatomy	0.7259
mmlu_astronomy	0.8750
mmlu_business_ethics	0.7400
mmlu_clinical_knowledge	0.7962
mmlu_college_biology	0.8542
mmlu_college_chemistry	0.5400
mmlu_college_computer_science	0.7100
mmlu_college_mathematics	0.5500
mmlu_college_medicine	0.7630
mmlu_college_physics	0.5588
mmlu_computer_security	0.8000
mmlu_conceptual_physics	0.8128
mmlu_econometrics	0.6579
mmlu_electrical_engineering	0.8069
mmlu_elementary_mathematics	0.6984
mmlu_formal_logic	0.6032
mmlu_global_facts	0.4100
mmlu_high_school_biology	0.9032
mmlu_high_school_chemistry	0.7586
mmlu_high_school_computer_science	0.8300
mmlu_high_school_european_history	0.8364
mmlu_high_school_geography	0.8788
mmlu_high_school_government_and_politics	0.9223
mmlu_high_school_macroeconomics	0.7795
mmlu_high_school_mathematics	0.4481
mmlu_high_school_microeconomics	0.8782
mmlu_high_school_physics	0.6623
mmlu_high_school_psychology	0.9101
mmlu_high_school_statistics	0.7454
mmlu_high_school_us_history	0.8431
mmlu_high_school_world_history	0.8608
mmlu_human_aging	0.7265
mmlu_human_sexuality	0.8626
mmlu_humanities	0.6408
mmlu_international_law	0.8430
mmlu_jurisprudence	0.8241
mmlu_logical_fallacies	0.8098
mmlu_machine_learning	0.5714
mmlu_management	0.8835
mmlu_marketing	0.9103
mmlu_medical_genetics	0.8700
mmlu_miscellaneous	0.8250
mmlu_moral_disputes	0.7197
mmlu_moral_scenarios	0.4391
mmlu_nutrition	0.8007
mmlu_other	0.7683
mmlu_philosophy	0.7331
mmlu_prehistory	0.8056
mmlu_professional_accounting	0.5993
mmlu_professional_law	0.5398
mmlu_professional_medicine	0.8382
mmlu_professional_psychology	0.7778
mmlu_public_relations	0.7364
mmlu_security_studies	0.7510
mmlu_social_sciences	0.8291
mmlu_sociology	0.8856
mmlu_stem	0.7165
mmlu_us_foreign_policy	0.8300
mmlu_virology	0.5301
mmlu_world_religions	0.8363
piqa	0.7650

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-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-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:

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.