Olmo-3.1-32B-Instruct-SFT-AutoRound-W4A16-Tuning

Model Details

This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of allenai/Olmo-3.1-32B-Instruct-SFT generated by TUNING. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	allenai/Olmo-3.1-32B-Instruct-SFT
Quantization Tool	TUNING
Quantization Scheme	W4A16
Quantized Size	17422 MB

Evaluation Results

Task	Accuracy
hellaswag	0.6119
mmlu	0.6864
mmlu_abstract_algebra	0.4900
mmlu_anatomy	0.5704
mmlu_astronomy	0.8289
mmlu_business_ethics	0.7200
mmlu_clinical_knowledge	0.7660
mmlu_college_biology	0.8611
mmlu_college_chemistry	0.4300
mmlu_college_computer_science	0.6900
mmlu_college_mathematics	0.5400
mmlu_college_medicine	0.6763
mmlu_college_physics	0.5686
mmlu_computer_security	0.7900
mmlu_conceptual_physics	0.7574
mmlu_econometrics	0.5965
mmlu_electrical_engineering	0.6759
mmlu_elementary_mathematics	0.6164
mmlu_formal_logic	0.5873
mmlu_global_facts	0.4500
mmlu_high_school_biology	0.8581
mmlu_high_school_chemistry	0.6404
mmlu_high_school_computer_science	0.8500
mmlu_high_school_european_history	0.7879
mmlu_high_school_geography	0.8535
mmlu_high_school_government_and_politics	0.9119
mmlu_high_school_macroeconomics	0.7462
mmlu_high_school_mathematics	0.4963
mmlu_high_school_microeconomics	0.8277
mmlu_high_school_physics	0.5894
mmlu_high_school_psychology	0.8936
mmlu_high_school_statistics	0.6944
mmlu_high_school_us_history	0.8382
mmlu_high_school_world_history	0.8523
mmlu_human_aging	0.7085
mmlu_human_sexuality	0.7481
mmlu_humanities	0.5915
mmlu_international_law	0.8017
mmlu_jurisprudence	0.8241
mmlu_logical_fallacies	0.8589
mmlu_machine_learning	0.6518
mmlu_management	0.8252
mmlu_marketing	0.8803
mmlu_medical_genetics	0.7700
mmlu_miscellaneous	0.8365
mmlu_moral_disputes	0.7543
mmlu_moral_scenarios	0.2927
mmlu_nutrition	0.8268
mmlu_other	0.7367
mmlu_philosophy	0.7396
mmlu_prehistory	0.7593
mmlu_professional_accounting	0.4894
mmlu_professional_law	0.4817
mmlu_professional_medicine	0.7096
mmlu_professional_psychology	0.7353
mmlu_public_relations	0.6455
mmlu_security_studies	0.7673
mmlu_social_sciences	0.7966
mmlu_sociology	0.8308
mmlu_stem	0.6708
mmlu_us_foreign_policy	0.8900
mmlu_virology	0.5241
mmlu_world_religions	0.8304
piqa	0.7867

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 = "Olmo-3.1-32B-Instruct-SFT-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 Olmo-3.1-32B-Instruct-SFT-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.