Ornith-1.0-35B-AutoRound-W4A16-Tuning

Model Details

This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of deepreinforce-ai/Ornith-1.0-35B generated by TUNING. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	deepreinforce-ai/Ornith-1.0-35B
Quantization Tool	TUNING
Quantization Scheme	W4A16
Quantized Size	19504 MB

Evaluation Results

Task	Accuracy
hellaswag	0.6320
mmlu	0.8221
mmlu_abstract_algebra	0.6600
mmlu_anatomy	0.8296
mmlu_astronomy	0.9276
mmlu_business_ethics	0.8700
mmlu_clinical_knowledge	0.8906
mmlu_college_biology	0.9375
mmlu_college_chemistry	0.6400
mmlu_college_computer_science	0.7800
mmlu_college_mathematics	0.6400
mmlu_college_medicine	0.8266
mmlu_college_physics	0.7549
mmlu_computer_security	0.8600
mmlu_conceptual_physics	0.9021
mmlu_econometrics	0.8070
mmlu_electrical_engineering	0.8069
mmlu_elementary_mathematics	0.8201
mmlu_formal_logic	0.6587
mmlu_global_facts	0.5100
mmlu_high_school_biology	0.9484
mmlu_high_school_chemistry	0.8079
mmlu_high_school_computer_science	0.9200
mmlu_high_school_european_history	0.8606
mmlu_high_school_geography	0.9394
mmlu_high_school_government_and_politics	0.9741
mmlu_high_school_macroeconomics	0.8923
mmlu_high_school_mathematics	0.6000
mmlu_high_school_microeconomics	0.9622
mmlu_high_school_physics	0.7616
mmlu_high_school_psychology	0.9596
mmlu_high_school_statistics	0.7963
mmlu_high_school_us_history	0.9265
mmlu_high_school_world_history	0.9241
mmlu_human_aging	0.8296
mmlu_human_sexuality	0.8855
mmlu_humanities	0.7552
mmlu_international_law	0.9008
mmlu_jurisprudence	0.8796
mmlu_logical_fallacies	0.9141
mmlu_machine_learning	0.7857
mmlu_management	0.8835
mmlu_marketing	0.9573
mmlu_medical_genetics	0.9400
mmlu_miscellaneous	0.9336
mmlu_moral_disputes	0.8410
mmlu_moral_scenarios	0.5654
mmlu_nutrition	0.8987
mmlu_other	0.8568
mmlu_philosophy	0.8682
mmlu_prehistory	0.8920
mmlu_professional_accounting	0.7199
mmlu_professional_law	0.6851
mmlu_professional_medicine	0.9154
mmlu_professional_psychology	0.8775
mmlu_public_relations	0.7273
mmlu_security_studies	0.8204
mmlu_social_sciences	0.9035
mmlu_sociology	0.9254
mmlu_stem	0.8084
mmlu_us_foreign_policy	0.9400
mmlu_virology	0.5602
mmlu_world_religions	0.9357
piqa	0.8101

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 = "Ornith-1.0-35B-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 Ornith-1.0-35B-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:

• Intel Neural Compressor
• AutoRound

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.