We've created a step-by-step guide to easily fine-tune from scratch with Kiln. No coding required.
The full guide is in our docs. It covers the following steps:- [2 mins]: Define task, goals, and schema
- [9 mins]: Synthetic data generation: create 920 high-quality examples for training
- [5 mins]: Dispatch 9 fine tuning jobs: Fireworks (Llama 3.2 1b/3b/11b, Llama 3.1 8b/70b, Mixtral 8x7b), OpenAI (GPT 4o, 4o-Mini), and Unsloth (Llama 3.2 1b/3b). Note: since this guide was written we've added fine-tuning on Together.ai, allowing tuning additional models like Qwen 2.5 14B/72B.
- [2 mins]: Deploy your new models and test they work
Once you've completed the guide, check out our evals guide for a strategy to evaluate your new models.
