Qwopus3.5-4B-Coder-AutoRound-W4A16-RTN

Model Details

This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of Jackrong/Qwopus3.5-4B-Coder generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	Jackrong/Qwopus3.5-4B-Coder
Quantization Tool	AutoRound
Quantization Scheme	W4A16
Original Size	1214 MB
Quantized Size	3271 MB

Evaluation Results

Task	Accuracy
hellaswag	0.5360
mmlu	0.7101
mmlu_abstract_algebra	0.5600
mmlu_anatomy	0.6741
mmlu_astronomy	0.8289
mmlu_business_ethics	0.7700
mmlu_clinical_knowledge	0.7811
mmlu_college_biology	0.8194
mmlu_college_chemistry	0.4800
mmlu_college_computer_science	0.6700
mmlu_college_mathematics	0.5000
mmlu_college_medicine	0.7457
mmlu_college_physics	0.4804
mmlu_computer_security	0.7600
mmlu_conceptual_physics	0.8043
mmlu_econometrics	0.6053
mmlu_electrical_engineering	0.7586
mmlu_elementary_mathematics	0.6825
mmlu_formal_logic	0.5238
mmlu_global_facts	0.4000
mmlu_high_school_biology	0.8935
mmlu_high_school_chemistry	0.7537
mmlu_high_school_computer_science	0.7900
mmlu_high_school_european_history	0.8121
mmlu_high_school_geography	0.8535
mmlu_high_school_government_and_politics	0.8912
mmlu_high_school_macroeconomics	0.7487
mmlu_high_school_mathematics	0.4889
mmlu_high_school_microeconomics	0.8697
mmlu_high_school_physics	0.6093
mmlu_high_school_psychology	0.8844
mmlu_high_school_statistics	0.6898
mmlu_high_school_us_history	0.8284
mmlu_high_school_world_history	0.8650
mmlu_human_aging	0.7085
mmlu_human_sexuality	0.8397
mmlu_humanities	0.6293
mmlu_international_law	0.8430
mmlu_jurisprudence	0.8333
mmlu_logical_fallacies	0.7975
mmlu_machine_learning	0.5446
mmlu_management	0.8350
mmlu_marketing	0.9145
mmlu_medical_genetics	0.8400
mmlu_miscellaneous	0.8161
mmlu_moral_disputes	0.7717
mmlu_moral_scenarios	0.4156
mmlu_nutrition	0.8137
mmlu_other	0.7570
mmlu_philosophy	0.7267
mmlu_prehistory	0.8025
mmlu_professional_accounting	0.5709
mmlu_professional_law	0.5228
mmlu_professional_medicine	0.7941
mmlu_professional_psychology	0.7418
mmlu_public_relations	0.7182
mmlu_security_studies	0.7510
mmlu_social_sciences	0.8050
mmlu_sociology	0.8607
mmlu_stem	0.6917
mmlu_us_foreign_policy	0.8600
mmlu_virology	0.5542
mmlu_world_religions	0.8070
piqa	0.7693

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 = "Qwopus3.5-4B-Coder-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 Qwopus3.5-4B-Coder-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.