the pretrained model was load many times.

[wang-tf]

Is there any reason load model weight for every subject_id loop ?

https://github.com/Fabian-Sc85/non-invasive-bp-estimation-using-deep-learning/blob/438bf3ff4aed9b1283effc85a50a769e790d5444/retrain_rppg_personalization.py#L126

[Fabian-Sc85]

Hi, yes this is intentional. We are retraining the pretrained neural network for every subject in our rPPG dataset independently which requires us to reload the pretrained model (or rather its weights) for every subject. This way we could investigate whether using subject-specific data (20 % of each subject‘s data) improves the prediction error after fine tuning.

[wang-tf]

Thanks, I understand.