This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of FlameF0X/Qwen2-0.2B-pt generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	FlameF0X/Qwen2-0.2B-pt
Quantization Tool	AutoRound
Quantization Scheme	W4A16
Quantized Size	309 MB

Evaluation Results

Task	Accuracy
hellaswag	0.2559
mmlu	0.2321
mmlu_abstract_algebra	0.2200
mmlu_anatomy	0.1778
mmlu_astronomy	0.1842
mmlu_business_ethics	0.2600
mmlu_clinical_knowledge	0.2151
mmlu_college_biology	0.2708
mmlu_college_chemistry	0.2200
mmlu_college_computer_science	0.2800
mmlu_college_mathematics	0.2100
mmlu_college_medicine	0.2023
mmlu_college_physics	0.2255
mmlu_computer_security	0.2800
mmlu_conceptual_physics	0.2681
mmlu_econometrics	0.2456
mmlu_electrical_engineering	0.2345
mmlu_elementary_mathematics	0.2063
mmlu_formal_logic	0.3095
mmlu_global_facts	0.1800
mmlu_high_school_biology	0.1839
mmlu_high_school_chemistry	0.1724
mmlu_high_school_computer_science	0.2600
mmlu_high_school_european_history	0.2242
mmlu_high_school_geography	0.2020
mmlu_high_school_government_and_politics	0.2124
mmlu_high_school_macroeconomics	0.2179
mmlu_high_school_mathematics	0.2111
mmlu_high_school_microeconomics	0.2437
mmlu_high_school_physics	0.1987
mmlu_high_school_psychology	0.2018
mmlu_high_school_statistics	0.1759
mmlu_high_school_us_history	0.2304
mmlu_high_school_world_history	0.2574
mmlu_human_aging	0.2870
mmlu_human_sexuality	0.2672
mmlu_humanities	0.2404
mmlu_international_law	0.2397
mmlu_jurisprudence	0.2315
mmlu_logical_fallacies	0.2454
mmlu_machine_learning	0.3036
mmlu_management	0.1845
mmlu_marketing	0.2949
mmlu_medical_genetics	0.3000
mmlu_miscellaneous	0.2388
mmlu_moral_disputes	0.2457
mmlu_moral_scenarios	0.2380
mmlu_nutrition	0.2255
mmlu_other	0.2388
mmlu_philosophy	0.1768
mmlu_prehistory	0.2130
mmlu_professional_accounting	0.2518
mmlu_professional_law	0.2464
mmlu_professional_medicine	0.1875
mmlu_professional_psychology	0.2582
mmlu_public_relations	0.1818
mmlu_security_studies	0.1918
mmlu_social_sciences	0.2272
mmlu_sociology	0.2388
mmlu_stem	0.2179
mmlu_us_foreign_policy	0.2900
mmlu_virology	0.2771
mmlu_world_religions	0.3099
piqa	0.5441

How to Use

HF Usage

Step 1: Install AutoRound

pip install auto-round

Step 2: Load and run the quantized model

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen2-0.2B-pt-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

bash

vllm serve Qwen2-0.2B-pt-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.

LeaderboardModel1/Qwen2-0.2B-pt-AutoRound-W4A16-RTN

Qwen2-0.2B-pt-AutoRound-W4A16-RTN

Model Details