burtenshaw/GemmaCoder3-12B

Model Card for burtenshaw/GemmaCoder-3-12B

This model is a fine-tuned version of google/gemma-3-12b-it on the open-r1/codeforces-cots dataset. It has been trained using TRL.

Evaluation

The model improves performance on the LiveCodeBench benchmark, but diminishes on other benchmarks.

Benchmark	GemmaCoder-12B	Gemma3-12B-it
Winogrande	63.9%	63.5%
MMLU	61.0%	69.5%
HellaSwag	54.0%	53.5%
LiveCodeBench	32.9%	21.9%

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="burtenshaw/gemma-3-12b-it-codeforces-SFT", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

Training procedure

This model was trained with SFT.

Framework versions

• TRL: 0.17.0.dev0
• Transformers: 4.51.0.dev0
• Pytorch: 2.6.0
• Datasets: 3.4.1
• 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}}
}