MagiSeek-Pro-V1-AutoRound-W4A16-Tuning

Model Details

This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of groxaxo/MagiSeek-Pro-V1 generated by TUNING. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	groxaxo/MagiSeek-Pro-V1
Quantization Tool	TUNING
Quantization Scheme	W4A16
Quantized Size	13575 MB

Evaluation Results

Task	Accuracy
hellaswag	0.6212
mmlu	0.7433
mmlu_abstract_algebra	0.4600
mmlu_anatomy	0.8074
mmlu_astronomy	0.8882
mmlu_business_ethics	0.7600
mmlu_clinical_knowledge	0.8151
mmlu_college_biology	0.8819
mmlu_college_chemistry	0.5400
mmlu_college_computer_science	0.7200
mmlu_college_mathematics	0.4900
mmlu_college_medicine	0.7457
mmlu_college_physics	0.6569
mmlu_computer_security	0.7900
mmlu_conceptual_physics	0.7745
mmlu_econometrics	0.5965
mmlu_electrical_engineering	0.7379
mmlu_elementary_mathematics	0.7434
mmlu_formal_logic	0.4603
mmlu_global_facts	0.5700
mmlu_high_school_biology	0.8871
mmlu_high_school_chemistry	0.7241
mmlu_high_school_computer_science	0.8600
mmlu_high_school_european_history	0.8242
mmlu_high_school_geography	0.8838
mmlu_high_school_government_and_politics	0.9689
mmlu_high_school_macroeconomics	0.7846
mmlu_high_school_mathematics	0.4926
mmlu_high_school_microeconomics	0.8403
mmlu_high_school_physics	0.5894
mmlu_high_school_psychology	0.9174
mmlu_high_school_statistics	0.6944
mmlu_high_school_us_history	0.8971
mmlu_high_school_world_history	0.8692
mmlu_human_aging	0.7982
mmlu_human_sexuality	0.8779
mmlu_humanities	0.6536
mmlu_international_law	0.8430
mmlu_jurisprudence	0.8241
mmlu_logical_fallacies	0.8589
mmlu_machine_learning	0.5804
mmlu_management	0.8835
mmlu_marketing	0.9444
mmlu_medical_genetics	0.8600
mmlu_miscellaneous	0.8991
mmlu_moral_disputes	0.7688
mmlu_moral_scenarios	0.4290
mmlu_nutrition	0.8562
mmlu_other	0.8069
mmlu_philosophy	0.7588
mmlu_prehistory	0.8364
mmlu_professional_accounting	0.6064
mmlu_professional_law	0.5587
mmlu_professional_medicine	0.8125
mmlu_professional_psychology	0.8007
mmlu_public_relations	0.7818
mmlu_security_studies	0.8408
mmlu_social_sciences	0.8460
mmlu_sociology	0.8905
mmlu_stem	0.7146
mmlu_us_foreign_policy	0.9100
mmlu_virology	0.5723
mmlu_world_religions	0.8596
piqa	0.8183

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 = "MagiSeek-Pro-V1-AutoRound-W4A16-Tuning"

# 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 MagiSeek-Pro-V1-AutoRound-W4A16-Tuning \
    --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.