README.md

---
base_model: barc0/Llama-3.1-ARC-Potpourri-Induction-8B
datasets:
- suehyunpark/induction_arc_inputs_train_bc_trajectories_max10-per-task_check_answer
library_name: transformers
model_name: barc0/Llama-3.1-ARC-Potpourri-Induction-8B
tags:
- generated_from_trainer
- alignment-handbook
- trl
- sft
licence: license
---

# Model Card for barc0/Llama-3.1-ARC-Potpourri-Induction-8B

This model is a fine-tuned version of [barc0/Llama-3.1-ARC-Potpourri-Induction-8B](https://huggingface.co/barc0/Llama-3.1-ARC-Potpourri-Induction-8B) on the [['suehyunpark/induction_arc_inputs_train_bc_trajectories_max10-per-task_check_answer']](https://huggingface.co/datasets/['suehyunpark/induction_arc_inputs_train_bc_trajectories_max10-per-task_check_answer']) dataset.
It has been trained using [TRL](https://github.com/huggingface/trl).

## Quick start

```python
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="suehyunpark/potpourri-8b-inst-fft-induction-bc-trajectory-max10-per-task-check-answer", 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"/>](https://wandb.ai/suehyun/arc-improve/runs/4y71yzaa)

This model was trained with SFT.

### Framework versions

- TRL: 0.12.1
- Transformers: 4.46.2
- Pytorch: 2.5.1
- Datasets: 3.1.0
- Tokenizers: 0.20.3

## Citations



Cite TRL as:
    
```bibtex
@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}}
}
```