SLEAP-NN Metrics Visualizer

Upload metrics.npz files to visualize and compare model performance

πŸ“Š
Drop metrics.npz or .json files here
Supports multiple files for comparison | Load sample data

πŸ“Š Summary Metrics

β–Ό

πŸ“ˆ Charts

β–Ό

Distance Error Distribution

PCK vs Threshold

Per-Keypoint Accuracy (mPCK)

Distance Error by Node

OKS Score Distribution

Visibility Confusion Matrix

🎯 Ground Truth Labels (required by Advanced Analytics · pose overlay · Visualize Frames)
🎞️ Video (shared, optional β€” enables image behind pose overlay)
⚠ Browsers can only decode .mp4, .mov, .webm, .m4v, and .mkv. .avi files won't load β€” convert first with:
uvx --from static-ffmpeg static_ffmpeg -i input.avi -c:v libx264 -pix_fmt yuv420p output.mp4
Video
+ Drop video file

πŸ”¬ Advanced Analytics

β–Ό
⚠ For the views below to match the numbers, each model's GT + preds SLPs must come from the same train/val split as its metrics NPZ. Use the per-model GT slot on each card if your models have different splits.
In Both (compare) mode this applies to every loaded model β€” otherwise single-model views are unaffected, and per-model errors stay correct for whichever model you select.

Per-Frame Timeline

Node Γ— Frame Heatmap

Worst Instances

Pose Overlay