I wish to create a exercise recognition with mediapipe and either PyTorch or TensorFlow but preferably the former

Gemini API
,
1

i need help on how to start and how to code

3

Hi @jianyao ,
You can take help from Gemini Models.
Thanks!