[Sample classification task] `azure_dataset_id` definition conflict between `GlaucomaPublicExt.py` and `GlaucomaPublic.py`

[JacopoTeneggi]

When following the sample classification task instructions to create the classification model configuration, we are instructed to specify azure_dataset_id in the constructor.

In the definition of GlaucomaPublic, though, these lines:

https://github.com/microsoft/InnerEye-DeepLearning/blob/eb5f931f200c253f8a3336fcb01cdaa58c6207be/InnerEye/ML/configs/classification/GlaucomaPublic.py#L12-L13

https://github.com/microsoft/InnerEye-DeepLearning/blob/eb5f931f200c253f8a3336fcb01cdaa58c6207be/InnerEye/ML/configs/classification/GlaucomaPublic.py#L23-L39

override the parameter, causing the experiment to fail if the user has not defined a dataset called glaucoma_public_dataset.

Expected behavior(s):

• The parameter is not overridden. or
• Documentation specifies to call the dataset glaucoma_public_dataset.

[ant0nsc]

Good catch, thanks for reporting! Update coming soon.