Post-Training Output
Audio Classification & Gesture Recognition
Trained model saved in the current project directory. Metrics and graphs retained for analysis. Model artifacts are ready for conversion and simulation via the Model Analysis tab.
Image Classification
Model files saved at: <project>/runs_yolo/train/weights/
| File | Description |
|---|---|
| best.pt | Checkpoint with best validation mAP50 — use this for inference and export |
| last.pt | Checkpoint from the final epoch |
The "Convert best.pt → ONNX" button in the Training Monitor tab converts the model for deployment (see Section 8 — Export & Deploy).