RichardErkhov/RyanYr_-_self-correct_Llama-3.2-3B-Instruct_MATH_0.5-0.75_bon_iter3-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

self-correctLlama-3.2-3B-InstructMATH0.5-0.75bon_iter3 - GGUF

• —Model creator: https://huggingface.co/RyanYr/
• —Original model: https://huggingface.co/RyanYr/self-correctLlama-3.2-3B-InstructMATH0.5-0.75bon_iter3/

Original model description: --- basemodel: RyanYr/self-correctLlama-3.2-3B-InstructMATH0.25-0.5boniter2 libraryname: transformers modelname: self-correctLlama-3.2-3B-InstructMATH0.5-0.75bon_iter3 tags:

• —generatedfromtrainer
• —trl
• —sft licence: license ---

Model Card for self-correctLlama-3.2-3B-InstructMATH0.5-0.75bon_iter3

This model is a fine-tuned version of RyanYr/self-correct_Llama-3.2-3B-Instruct_MATH_0.25-0.5_bon_iter2. It has been trained using TRL.

Quick start

from transformers import pipeline

question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="RyanYr/self-correct_Llama-3.2-3B-Instruct_MATH_0.5-0.75_bon_iter3", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

Training procedure

<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>

This model was trained with SFT.

Framework versions

• —TRL: 0.12.0.dev0
• —Transformers: 4.45.2
• —Pytorch: 2.4.0
• —Datasets: 3.0.1
• —Tokenizers: 0.20.1

Citations

Cite TRL as:

@misc{vonwerra2022trl,
	title        = {{TRL: Transformer Reinforcement Learning}},
	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
	year         = 2020,
	journal      = {GitHub repository},
	publisher    = {GitHub},
	howpublished = {\url{https://github.com/huggingface/trl}}
}

Additional thanks to @nicoboss for giving me access to his private supercomputer, enabling me to provide many more quants, at much higher speed, than I would otherwise be able to.