search

mist-medical / MIST

MIST: A simple, scalable, and end-to-end framework for 3D medical imaging segmentation.
Apache License 2.0
21 stars 12 forks source link

Old training data not removed when changing dataset #4

Closed rachelglenn closed 7 months ago

rachelglenn commented 1 year ago

When I switch datasets, the preprocessing adds the new dataset to the old dataset directories numpy/images numpy/labels

If we can remove these directories to avoid this issue, when running the preprocessing, it would really help.

https://github.com/aecelaya/MIST/blob/9824ba8e6ff728586c14afe4e26992426a7e737b/mist-torch/runtime/utils.py#L30

def create_empty_dir(path): if not (os.path.exists(path)): os.makedirs(path) else: os.system("rm -rf " + path) os.makedirs(path)

aecelaya commented 1 year ago

Can you give me a specific example of this please? When running on a new dataset, you can specify the directory where the preprocessed data is saved. Retaining the preprocessed data for a given dataset is useful when you want to run multiple iterations of training with different models and/or loss functions.