Compare Virtual Staining Models with Metrics #11
Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
There was a problem hiding this comment.
Nice job! I left a few comments throughout this PR. Generally I focused on the software implementation, I don't have as much insight into the modeling efforts (where you may want to get another set of eyes).
In addition to what I mentioned in the comments I strongly urge for there to be software tests added to this work to help better understand or maintain the modules and classes found in this PR. Generally as code is added it's a good idea to add tests to help ensure there are no unforeseen consequences of what you're developing. That said, it could be that you iteratively add tests external to this PR.
|
|
||
|
|
||
| class ImageMetaDataset(Dataset): | ||
| """Iterable Image Dataset for Stained Images, which supports applying transformations to the inputs and targets""" |
There was a problem hiding this comment.
This line looks a little longer than the usual Python line length expectations. Consider using multiple lines and adding a formatter to check for this type of thing (Black or ruff-format, for example).
There was a problem hiding this comment.
Good idea, I'll change that
This pr focuses on showing virtual stain model performances with four metrics across the best models thus far. The models evaluated were determined to be the best models using their respective generator losses. Any constructive feedback is welcome while reviewing this pr.