Terumo: binary classifiers for glomerular data

• Usage:

1. Copy all the dataset folders to the data/raw folder.
2. Run make build_dataset. This will run in two steps:
   1. First, it will arrange the dataset into 6 sub-folders of the data/raw folder, according to class names:
      1. Hypercellularity
      2. Membranous
      3. Sclerosis
      4. Normal
      5. Podocytopathy
      6. Crescent
   2. Then, it will create the 6 copies that will be used to train each binary classifier
3. (Optional) If you want a test run (of 1 epoch, and only a single model) run make test
4. Run make train to train all 6 models sequentially, in the order above (Hypercellularity, Membranous, Sclerosis, Normal, Podocytopathy, Crescent)