Estimates heart rate (bpm) from a wrist PPG signal + 3-axis accelerometer, using a small 1D CNN trained on the public PPG-DaLiA dataset. Wearable sensor data → actionable health metric.
/predict — one 8 s window (PPG 512@64Hz + ACC 256×3@32Hz) → HR in bpm/health — status + model metadata/docs — interactive API explorercurl -X POST .../predict -H "Content-Type: application/json" -d @example_payload.json
Validation MAE ≈ 12 bpm (held-out subjects); naive baseline ≈ 24 bpm.