Magidonia-24B-v4.3-creative-ORPO-AutoRound-W4A16-RTN

Model Details

This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of WarlordHermes/Magidonia-24B-v4.3-creative-ORPO generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	WarlordHermes/Magidonia-24B-v4.3-creative-ORPO
Quantization Tool	AutoRound
Quantization Scheme	W4A16
Quantized Size	13575 MB

Evaluation Results

Task	Accuracy
hellaswag	0.6524
mmlu	0.7631
mmlu_abstract_algebra	0.5500
mmlu_anatomy	0.8222
mmlu_astronomy	0.8684
mmlu_business_ethics	0.7700
mmlu_clinical_knowledge	0.8491
mmlu_college_biology	0.8958
mmlu_college_chemistry	0.5400
mmlu_college_computer_science	0.7100
mmlu_college_mathematics	0.6100
mmlu_college_medicine	0.7688
mmlu_college_physics	0.6471
mmlu_computer_security	0.8300
mmlu_conceptual_physics	0.8128
mmlu_econometrics	0.6316
mmlu_electrical_engineering	0.7517
mmlu_elementary_mathematics	0.7937
mmlu_formal_logic	0.5317
mmlu_global_facts	0.6700
mmlu_high_school_biology	0.8903
mmlu_high_school_chemistry	0.7192
mmlu_high_school_computer_science	0.9000
mmlu_high_school_european_history	0.8364
mmlu_high_school_geography	0.8889
mmlu_high_school_government_and_politics	0.9741
mmlu_high_school_macroeconomics	0.7872
mmlu_high_school_mathematics	0.5519
mmlu_high_school_microeconomics	0.8571
mmlu_high_school_physics	0.6358
mmlu_high_school_psychology	0.9174
mmlu_high_school_statistics	0.7222
mmlu_high_school_us_history	0.9020
mmlu_high_school_world_history	0.8903
mmlu_human_aging	0.7758
mmlu_human_sexuality	0.8626
mmlu_humanities	0.6797
mmlu_international_law	0.8760
mmlu_jurisprudence	0.8704
mmlu_logical_fallacies	0.8160
mmlu_machine_learning	0.7054
mmlu_management	0.8835
mmlu_marketing	0.9359
mmlu_medical_genetics	0.8900
mmlu_miscellaneous	0.9004
mmlu_moral_disputes	0.7832
mmlu_moral_scenarios	0.4715
mmlu_nutrition	0.8627
mmlu_other	0.8169
mmlu_philosophy	0.7749
mmlu_prehistory	0.8395
mmlu_professional_accounting	0.6064
mmlu_professional_law	0.5939
mmlu_professional_medicine	0.8529
mmlu_professional_psychology	0.8366
mmlu_public_relations	0.7455
mmlu_security_studies	0.8082
mmlu_social_sciences	0.8531
mmlu_sociology	0.8856
mmlu_stem	0.7466
mmlu_us_foreign_policy	0.9500
mmlu_virology	0.5542
mmlu_world_religions	0.8655
piqa	0.8030

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 = "Magidonia-24B-v4.3-creative-ORPO-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 Magidonia-24B-v4.3-creative-ORPO-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.