-
The code in tumor_segmentation/main.py (https://github.com/neuropoly/ivado-medical-imaging/blob/al/tumor_segmentation_UNet2D/tumor_segmentation/main.py) is essentially the same as ivadomed/main.py (ht…
-
The base U-Net currently takes ~600 seconds per epoch to train, while FiLMed U-Net takes ~340s. This makes no sense since FiLM adds complexity to the model, thus it should take longer to train (which …
-
See comment: [comment](https://github.com/neuropoly/ivado-medical-imaging/pull/28#issuecomment-505562057)
We also need to add `matplotlib` as requirement, see [comment](https://github.com/neuropoly…