CoolFace
Modelpublic

LaTexT/qwen3-8b-gist-sft-50k-gz7-newlines

sourceHugging Faceupdated 14d agoView on Hugging Face
0likes198downloads
Model Card

Model Card for None

This model is a fine-tuned version of Qwen/Qwen3-8B. It has been trained using 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="None", 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
  • —Transformers: 4.51.1
  • —Pytorch: 2.5.1+cu124
  • —Datasets: 3.6.0
  • —Tokenizers: 0.21.1

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}}
}

Provenance (FAIR cluster backup, 2026-09-14)

  • —Source dir: /checkpoint/comem/shannons/latent-cot/ckpts-w32/20260102-004613~gist_sft~Qwen3-8B~ot3-1.2m-50k~bs128~lr4e-5~epoch5~l18000~gz7+d-newlines+cross_gist+input+wrap_gist_token+wrap_gist_token_to_special_tokens
  • —Launch: train/commands/week32/ (gist_sft 50k)
  • —Base model: Qwen/Qwen3-8B · Dataset: shannons/ot3-1.2m-50k
  • —Paper row: Paper Table 2: gist_sft 50k g7-newlines