Mythos-nano-gguf-free-AutoRound-NVFP4-Tuning

Model Details

This model is a NVFP4 (NVIDIA FP4) quantization of FlameF0X/Mythos-nano-gguf-free generated by TUNING. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	FlameF0X/Mythos-nano-gguf-free
Quantization Tool	TUNING
Quantization Scheme	NVFP4
Quantized Size	2548 MB

Evaluation Results

Task	Accuracy
hellaswag	0.2642
mmlu	0.2296
mmlu_abstract_algebra	0.2000
mmlu_anatomy	0.1926
mmlu_astronomy	0.1776
mmlu_business_ethics	0.3100
mmlu_clinical_knowledge	0.2226
mmlu_college_biology	0.2708
mmlu_college_chemistry	0.1900
mmlu_college_computer_science	0.2300
mmlu_college_mathematics	0.2200
mmlu_college_medicine	0.1965
mmlu_college_physics	0.1961
mmlu_computer_security	0.2800
mmlu_conceptual_physics	0.2511
mmlu_econometrics	0.2456
mmlu_electrical_engineering	0.2414
mmlu_elementary_mathematics	0.2143
mmlu_formal_logic	0.2937
mmlu_global_facts	0.1700
mmlu_high_school_biology	0.1677
mmlu_high_school_chemistry	0.1626
mmlu_high_school_computer_science	0.2500
mmlu_high_school_european_history	0.2121
mmlu_high_school_geography	0.1869
mmlu_high_school_government_and_politics	0.1969
mmlu_high_school_macroeconomics	0.2205
mmlu_high_school_mathematics	0.2037
mmlu_high_school_microeconomics	0.2185
mmlu_high_school_physics	0.1987
mmlu_high_school_psychology	0.1908
mmlu_high_school_statistics	0.1528
mmlu_high_school_us_history	0.2549
mmlu_high_school_world_history	0.2616
mmlu_human_aging	0.3049
mmlu_human_sexuality	0.2519
mmlu_humanities	0.2423
mmlu_international_law	0.2397
mmlu_jurisprudence	0.2593
mmlu_logical_fallacies	0.2209
mmlu_machine_learning	0.2946
mmlu_management	0.1456
mmlu_marketing	0.2906
mmlu_medical_genetics	0.3100
mmlu_miscellaneous	0.2350
mmlu_moral_disputes	0.2543
mmlu_moral_scenarios	0.2380
mmlu_nutrition	0.2353
mmlu_other	0.2382
mmlu_philosophy	0.2154
mmlu_prehistory	0.2130
mmlu_professional_accounting	0.2376
mmlu_professional_law	0.2438
mmlu_professional_medicine	0.1801
mmlu_professional_psychology	0.2467
mmlu_public_relations	0.2909
mmlu_security_studies	0.1755
mmlu_social_sciences	0.2223
mmlu_sociology	0.2687
mmlu_stem	0.2093
mmlu_us_foreign_policy	0.2600
mmlu_virology	0.2711
mmlu_world_religions	0.2924
piqa	0.5218

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 = "Mythos-nano-gguf-free-AutoRound-NVFP4-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 Mythos-nano-gguf-free-AutoRound-NVFP4-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.