Tess-4-35B-A3B-AutoRound-W4A16-RTN

Model Details

This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of Don-oz/Tess-4-35B-A3B generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	Don-oz/Tess-4-35B-A3B
Quantization Tool	AutoRound
Quantization Scheme	W4A16
Quantized Size	19936 MB

Evaluation Results

Task	Accuracy
hellaswag	0.6321
mmlu	0.8210
mmlu_abstract_algebra	0.5900
mmlu_anatomy	0.8519
mmlu_astronomy	0.9342
mmlu_business_ethics	0.8400
mmlu_clinical_knowledge	0.8981
mmlu_college_biology	0.9444
mmlu_college_chemistry	0.6400
mmlu_college_computer_science	0.7800
mmlu_college_mathematics	0.6300
mmlu_college_medicine	0.8439
mmlu_college_physics	0.6078
mmlu_computer_security	0.8800
mmlu_conceptual_physics	0.9404
mmlu_econometrics	0.7982
mmlu_electrical_engineering	0.8276
mmlu_elementary_mathematics	0.7857
mmlu_formal_logic	0.6825
mmlu_global_facts	0.4900
mmlu_high_school_biology	0.9484
mmlu_high_school_chemistry	0.7980
mmlu_high_school_computer_science	0.8800
mmlu_high_school_european_history	0.8606
mmlu_high_school_geography	0.9242
mmlu_high_school_government_and_politics	0.9845
mmlu_high_school_macroeconomics	0.8821
mmlu_high_school_mathematics	0.6074
mmlu_high_school_microeconomics	0.9580
mmlu_high_school_physics	0.8013
mmlu_high_school_psychology	0.9523
mmlu_high_school_statistics	0.7824
mmlu_high_school_us_history	0.9069
mmlu_high_school_world_history	0.9114
mmlu_human_aging	0.8072
mmlu_human_sexuality	0.8855
mmlu_humanities	0.7566
mmlu_international_law	0.9174
mmlu_jurisprudence	0.8796
mmlu_logical_fallacies	0.9141
mmlu_machine_learning	0.8036
mmlu_management	0.9126
mmlu_marketing	0.9402
mmlu_medical_genetics	0.9300
mmlu_miscellaneous	0.9387
mmlu_moral_disputes	0.8468
mmlu_moral_scenarios	0.5899
mmlu_nutrition	0.8889
mmlu_other	0.8565
mmlu_philosophy	0.8650
mmlu_prehistory	0.9167
mmlu_professional_accounting	0.7128
mmlu_professional_law	0.6741
mmlu_professional_medicine	0.9301
mmlu_professional_psychology	0.8758
mmlu_public_relations	0.7273
mmlu_security_studies	0.8408
mmlu_social_sciences	0.9015
mmlu_sociology	0.9303
mmlu_stem	0.8034
mmlu_us_foreign_policy	0.9400
mmlu_virology	0.5783
mmlu_world_religions	0.9064
piqa	0.8221

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 = "Tess-4-35B-A3B-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 Tess-4-35B-A3B-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.