TL;DR
Over 250 notebooks from Unsloth for fine-tuning and reinforcement learning (e.g., GRPO) of text, vision, audio, embedding, and TTS models, all runnable directly on Colab.
Key features
Wide model support: Notebooks for fine-tuning latest models like Gemma 4, Qwen3.5, Qwen3, gpt-oss, etc.
Multimodal: Covers text, vision, audio, embedding, TTS modalities.
Reinforcement learning: Includes RL algorithms like GRPO (Group Relative Policy Optimization).
Instant execution: All notebooks run with one click on Colab.
Complete pipeline: Data preparation, training, and inference in a single notebook.
When to use it
When you want to quickly fine-tune or apply reinforcement learning to a specific model.
When you want to experiment with the latest fine-tuning techniques on Colab without additional setup.
When you need to manage models across multiple modalities (text, vision, audio, etc.) in one place.