​
What is a Snapshot?

A snapshot is a saved version of your dataset at a specific point in time. It is useful to keep track of your dataset’s progress and to revert to a previous state if needed.

​
How to Create a Snapshot

Need at least 10 datapoints in training to create a snapshot. You can either use autosplit or manually split your dataset.
1

Click Snapshots 📸

Click the Snapshots button on the top right of the screen to create a snapshot of your dataset.GIF of clicking the snapshots button
2

Enter a Name for the Snapshot

Create a name for the snapshot. You’ll need it when you want to evaluate or fine-tune a model.GIF of entering a name for the snapshot
3

Download the Snapshot

Click the download button to download the snapshot. Save it somewhere you can remember to access it later.GIF of downloading a snapshot
4

Fine-Tune or Evaluate with the Snapshot

A Snapshot is required to fine-tune or evaluate a model.

Fine-Tune

Fine-Tune a model with a snapshot.

Evaluate

Evaluate a model with a snapshot.