There seems to be an issue with the Gesture Dataset, having repetitive channels. The below shows the training dataset. <img width="686" alt="스크린샷 2024-02-28 오후 2 38 25" src="https://github.com/mims-harvard/TFC-pretraining/assets/38973453/8b2049e5-a2e4-4dff-9333-88f9616d1df1">