> ## Documentation Index > Fetch the complete documentation index at: https://docs.baseten.co/llms.txt > Use this file to discover all available pages before exploring further. # Choose SFT or RL > Choose supervised fine-tuning when you have answers to imitate and reinforcement learning when you can score outcomes, and combine them when a task needs both. Both methods train a LoRA adapter on Loops. They differ in what you provide. Supervised fine-tuning (SFT) needs examples of the right answer. Reinforcement learning (RL) needs a way to score an attempt. Choose by which one you can provide reliably. | | SFT | RL | | - | - | - | | You provide | Conversations whose assistant turns are the behavior you want | Tasks, an agent that attempts them, and a grader that scores each attempt | | The model learns | To reproduce the targets | Which of its own attempts scored better | | Quality ceiling | The quality of your targets | What your grader rewards | | Fits tasks where | You can write or collect correct outputs, such as extraction, classification, formatting, and summarization | You can check a result more easily than write it, such as code that must pass tests or an agent that must leave a system in a given state | | Main risk | Learning the mistakes in the data | Learning to satisfy the grader without doing the task | | Start with | [Prepare SFT data](https://docs.baseten.co/loops/sft-data.md) | [Design RL environments and rewards](https://docs.baseten.co/loops/rl-environments.md) | ## Choose SFT when you have answers SFT is the simpler and cheaper method, and usually the first to try. It fits when correct outputs exist or can be produced: production outputs you've verified, corrections from reviewers, or outputs from a stronger model that pass your evaluation. Distilling a large model's behavior on one task into a smaller model is SFT on the large model's filtered outputs. SFT can't exceed its targets. If your best outputs come from production, the trained model learns production's behavior, mistakes included. See [choose targets you trust](https://docs.baseten.co/loops/sft-data.md#choose-targets-you-trust). ## Choose RL when you can score outcomes RL fits when checking a result is easier than writing it. Unit tests can check code no one has written yet. A shell command can check that an agent left the files in the right state. A rubric can check a summary against its source. RL also trains on the model's own attempts, so it learns from the situations the model actually gets into, such as a failed tool call in the middle of a long task. RL needs a starting model that sometimes succeeds. The model learns from the difference between its better and worse attempts at the same task, so a task it always fails teaches nothing. See [size tasks so attempts disagree](https://docs.baseten.co/loops/rl-environments.md#size-tasks-so-attempts-disagree). ## Combine them A common sequence is SFT first, then RL. SFT teaches the format and the basic behavior from examples, so the model succeeds often enough for RL to have a signal. RL then improves on the cases where the examples ran out. In the Loops cookbook, start the RL run from the SFT run's trainer weights with `[training] init_from = "bt://loops:/weights/step-"`. The base model and LoRA rank must match the SFT run. If SFT plateaus below your bar, look at what kind of errors remain before you add RL. Errors that come from the data, such as inconsistent labels, need better data. Errors the model makes in situations the data never showed, such as recovering from a failed step, are what RL trains on. ## Next steps * [Fine-tune tool calling with SFT](https://docs.baseten.co/examples/sft-tool-calling.md): An SFT run with a before-and-after evaluation. * [Train a model to count circles with RL](https://docs.baseten.co/examples/rl-counting.md): A first RL run on serverless Loops with an image generator and a correctness reward. * [Train a tool-using agent with RL](https://docs.baseten.co/examples/rl-tool-agent.md): Build a custom environment with a sandbox and a hidden grader. * [Ship a post-trained model](https://docs.baseten.co/post-training/ship.md): Check, compare, deploy, and roll out the result.