9. What's next¶

You've been through the whole workflow: label, train, correct, track, proofread and export. You're done with the tutorial!

Improve and reuse your model¶

Label more frames and retrain. Each round of correcting and retraining makes the model better.
Run it on more videos, as you did in step 6, or all at once with Inference Target: All videos. Proofread the tracks of each before you analyze them.
Explore the app. The Guides cover everything it can do.

SLEAP gives you where each body part is on every frame. Two common next steps:

Behavior classification: label behaviors frame by frame from the pose trajectories, for example with Keypoint-MoSeq (unsupervised) or SimBA (supervised).
Movement analysis: measure speed, posture and coordination from the trajectories, for example with movement.

For data loading and I/O in Python, use sleap-io.

The SLEAP docs keep a longer list of analysis tools, example notebooks and reading by research area: I'm done SLEAPing, now what?