tunix

Here are 3 public repositories matching this topic...

Training Gemma3-1B to produce structured reasoning traces using Tunix (SFT + GRPO) on Kaggle TPUs. Google Tunix Hackathon submission.

machine-learning reinforcement-learning kaggle gemma reasoning llm chain-of-thought tunix

Fine-tuning SLM (Gemma 3 1B) with Tunix, to create a "reasoning" model, that explains its Chain-of-Thought for the Mortgage Underwriting process.

docker google web-app kaggle-competition slm sft grpo gemma3-1b tunix mortgage-underwriter

Teaching Gemma2 to reason transparently via multi-domain GRPO training | 100% format accuracy, 60.8% exact accuracy | Google Tunix Hackathon 2025

google hackathon gemma reasoning tpu jax llm grpo tunix

Add a description, image, and links to the tunix topic page so that developers can more easily learn about it.

To associate your repository with the tunix topic, visit your repo's landing page and select "manage topics."