Skip to content

Refactor data loading in `train.py` for memory efficiency

PAUL M CURRY requested to merge pcurry/improvements into main
  • Replace lists of volumes and labels with DataFrames for efficient concatenation.
  • Update loop to process each NIfTI file individually, freeing memory after processing.
  • Use index of file in loop to aide in processing.
  • Modify how the source file is tracked within the DataFrame.
  • Correct plotting function to use a single sample volume instead of combined volumes.
  • Ensure proper handling of variables for most important segment visualization.

Merge request reports