Current patient sampling methods at the patient-level are limited and could benefit from integration with PyTorch samplers for more flexibility.
Proposed Solution
Integration of PyTorch Samplers: Allow the use of PyTorch's built-in samplers to facilitate more dynamic and statistically robust patient sampling methods.
Update Class Structures: Remove the patient_sampling manual implementation.
Enhance Patient Sampling Using PyTorch Samplers
Problem
Current patient sampling methods at the patient-level are limited and could benefit from integration with PyTorch samplers for more flexibility.
Proposed Solution