This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of llmfan46/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	llmfan46/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic
Quantization Tool	AutoRound
Quantization Scheme	W4A16
Quantized Size	14655 MB

Evaluation Results

Task	Accuracy
hellaswag	0.3486
mmlu	0.5051
mmlu_abstract_algebra	0.3900
mmlu_anatomy	0.3926
mmlu_astronomy	0.4539
mmlu_business_ethics	0.5100
mmlu_clinical_knowledge	0.4792
mmlu_college_biology	0.5972
mmlu_college_chemistry	0.4100
mmlu_college_computer_science	0.6000
mmlu_college_mathematics	0.3600
mmlu_college_medicine	0.4451
mmlu_college_physics	0.4020
mmlu_computer_security	0.4600
mmlu_conceptual_physics	0.4511
mmlu_econometrics	0.3860
mmlu_electrical_engineering	0.4069
mmlu_elementary_mathematics	0.4656
mmlu_formal_logic	0.4524
mmlu_global_facts	0.2900
mmlu_high_school_biology	0.6000
mmlu_high_school_chemistry	0.4236
mmlu_high_school_computer_science	0.6100
mmlu_high_school_european_history	0.7576
mmlu_high_school_geography	0.6515
mmlu_high_school_government_and_politics	0.7668
mmlu_high_school_macroeconomics	0.5179
mmlu_high_school_mathematics	0.3481
mmlu_high_school_microeconomics	0.6176
mmlu_high_school_physics	0.4106
mmlu_high_school_psychology	0.6569
mmlu_high_school_statistics	0.5185
mmlu_high_school_us_history	0.8088
mmlu_high_school_world_history	0.7764
mmlu_human_aging	0.4978
mmlu_human_sexuality	0.4962
mmlu_humanities	0.5209
mmlu_international_law	0.6529
mmlu_jurisprudence	0.4444
mmlu_logical_fallacies	0.5092
mmlu_machine_learning	0.5357
mmlu_management	0.4078
mmlu_marketing	0.4872
mmlu_medical_genetics	0.5200
mmlu_miscellaneous	0.3653
mmlu_moral_disputes	0.4364
mmlu_moral_scenarios	0.4358
mmlu_nutrition	0.4869
mmlu_other	0.4506
mmlu_philosophy	0.4437
mmlu_prehistory	0.5741
mmlu_professional_accounting	0.4007
mmlu_professional_law	0.4883
mmlu_professional_medicine	0.6728
mmlu_professional_psychology	0.5147
mmlu_public_relations	0.4364
mmlu_security_studies	0.5878
mmlu_social_sciences	0.5749
mmlu_sociology	0.5373
mmlu_stem	0.4672
mmlu_us_foreign_policy	0.6100
mmlu_virology	0.3976
mmlu_world_religions	0.5614
piqa	0.5800

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 = "gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-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 gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-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/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-AutoRound-W4A16-RTN

gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-AutoRound-W4A16-RTN

Model Details