Qwable-v1-AutoRound-W4A16-RTN

Model Details

This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of lordx64/Qwable-v1 generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	lordx64/Qwable-v1
Quantization Tool	AutoRound
Quantization Scheme	W4A16
Quantized Size	19936 MB

Evaluation Results

Task	Accuracy
hellaswag	0.6308
mmlu	0.8169
mmlu_abstract_algebra	0.6100
mmlu_anatomy	0.8444
mmlu_astronomy	0.9276
mmlu_business_ethics	0.8500
mmlu_clinical_knowledge	0.8830
mmlu_college_biology	0.9375
mmlu_college_chemistry	0.6400
mmlu_college_computer_science	0.7500
mmlu_college_mathematics	0.6500
mmlu_college_medicine	0.8150
mmlu_college_physics	0.6373
mmlu_computer_security	0.8700
mmlu_conceptual_physics	0.9362
mmlu_econometrics	0.7632
mmlu_electrical_engineering	0.8069
mmlu_elementary_mathematics	0.7884
mmlu_formal_logic	0.7063
mmlu_global_facts	0.5500
mmlu_high_school_biology	0.9516
mmlu_high_school_chemistry	0.7931
mmlu_high_school_computer_science	0.8900
mmlu_high_school_european_history	0.8424
mmlu_high_school_geography	0.9394
mmlu_high_school_government_and_politics	0.9793
mmlu_high_school_macroeconomics	0.8795
mmlu_high_school_mathematics	0.5926
mmlu_high_school_microeconomics	0.9622
mmlu_high_school_physics	0.7881
mmlu_high_school_psychology	0.9541
mmlu_high_school_statistics	0.7546
mmlu_high_school_us_history	0.9216
mmlu_high_school_world_history	0.9114
mmlu_human_aging	0.8072
mmlu_human_sexuality	0.8931
mmlu_humanities	0.7513
mmlu_international_law	0.9008
mmlu_jurisprudence	0.8889
mmlu_logical_fallacies	0.8834
mmlu_machine_learning	0.7589
mmlu_management	0.9320
mmlu_marketing	0.9402
mmlu_medical_genetics	0.9200
mmlu_miscellaneous	0.9361
mmlu_moral_disputes	0.8439
mmlu_moral_scenarios	0.5911
mmlu_nutrition	0.8791
mmlu_other	0.8532
mmlu_philosophy	0.8714
mmlu_prehistory	0.8951
mmlu_professional_accounting	0.7057
mmlu_professional_law	0.6623
mmlu_professional_medicine	0.9338
mmlu_professional_psychology	0.8725
mmlu_public_relations	0.7455
mmlu_security_studies	0.8286
mmlu_social_sciences	0.9006
mmlu_sociology	0.9303
mmlu_stem	0.7973
mmlu_us_foreign_policy	0.9400
mmlu_virology	0.5602
mmlu_world_religions	0.9123
piqa	0.8210

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 = "Qwable-v1-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 Qwable-v1-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.