CHAP2.0

CHAP2.0 contains PyTorch code for applying and fine-tuning CHAP models on wrist and hip accelerometer posture classification tasks, including iWatch and SOL/PASOS workflows.

The main supported workflows are:

• CHAP-ZS: zero-shot prediction using CHAP checkpoints.
• CHAP-FT: fine-tuning CHAP checkpoints on a target wrist or hip dataset.

---

Table of contents

• Getting Started
• Datasets and Models
• Advanced Usages