# Stop Training Early

Early Stopping lets you stop a training job while the model is training.

You may want to use this feature if:

1. Your training graphs show strong model performance and;
2. There are still many epochs left in your training job.

When you stop a model training early, the weights will be saved and the model will be available for use.

If you want to fully cancel a job for any reason, refer to the [Cancel a Training Job documentation](/train/cancel-a-training-job.md). Cancelled jobs do not save the model weights from training.

To stop a model training job early, click the "Stop Training Early" button:

<figure><img src="/files/pygrBLgl8sO8l0qqDm6u" alt=""><figcaption></figcaption></figure>

When you click the "Stop Training Early" button, work will immediately begin to stop the training job and prepare the model weights for use.

A tag will appear on your training job that denotes that the job has been stopped early:

<figure><img src="/files/ZAYo858f8bcAeTWSprVw" alt=""><figcaption></figcaption></figure>

It may take several minutes for a stopped model to be available.

When your model is ready, there will be a green checkmark next to the model version name:

<figure><img src="/files/yF25iu1F0FpsDu5hwFKN" alt=""><figcaption></figcaption></figure>

## NAS Training Stop Reasons

For models trained with Neural Architecture Search (NAS), training may also stop automatically when the model converges (reaches peak performance before all target epochs complete). In this case, a banner on the training results page will indicate that training converged and unnecessary epochs were skipped.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.roboflow.com/train/stop-training-early.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.