hussamalafandi/smollm2-sft-rewrite

Model Card for smollm2-sft-results

This model is a fine-tuned version of HuggingFaceTB/SmolLM2-135M. It has been trained using TRL.

Quick start

from transformers import pipeline

prompt = [
    {'content': "You're an AI assistant for text re-writing. Rewrite the input "
     'text to make it more concise while preserving its core meaning.',
     'role': 'system'},

    {'content': 'Hey Alex,\n'
     '\n'
     "I hope you're doing well! It's been a while since we met at the "
     'film festival last year. I was the one with the short film about '
     "the old abandoned factory. Anyway, I'm reaching out because I'm "
     'currently working on my thesis film project and I could really '
     'use some advice on cinematography. I remember our conversation '
     'about visual storytelling and I was hoping you might have some '
     'tips or insights to share.\n'
     '\n'
     'My film is a drama set in a small town, and I want to capture '
     'the mood and atmosphere of the location through my '
     "cinematography. I'm planning to shoot on location next month, "
     "but I'm still trying to figure out the best way to approach it. "
     'If you have any suggestions or resources you could point me to, '
     'I would be incredibly grateful.\n'
     '\n'
     "Also, I heard from a mutual friend that you're having a "
     'photography exhibition soon. Congratulations! I would love to '
     "attend if you don't mind sending me the details.\n"
     '\n'
     'Thanks in advance for any help you can provide. I really '
     'appreciate it.\n'
     '\n'
     'Best,\n'
     'Jordan',
     'role': 'user'}]

generator = pipeline("text-generation", model="hussamalafandi/smollm2-sft-rewrite", device="cuda")

output = generator(prompt, max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

Training procedure

This model was trained with SFT.

Framework versions

• TRL: 0.16.1
• Transformers: 4.51.3
• Pytorch: 2.6.0+cu124
• Datasets: 3.5.0
• Tokenizers: 0.21.1

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}}
}