TorchIO tutorial.ipynb

[idhamari]

Thank you for your effort. The tutorial is simple, clear and run without issue.
Could you please add a section to the tutorial that uses the model in production e.g.

1. Reading a single image and transforming it (without using dataset class).
2. Applying the inference using a pre-trained model to get a result image and save it with the original size.
3. Compute the dice of the original label (without resizing) and the saved output from step2.

I think there will be an issue as the used transforms do not preserve the spacing and directions.