VCP tutorials #227
VCP tutorials #227
Conversation
ziw-liu
added
documentation
| # %% | ||
| # Install VisCy with the optional dependencies for this example | ||
| # See the [repository](https://github.com/mehta-lab/VisCy) for more details | ||
| # !pip install "viscy[metrics,visual]==0.3.0rc2" |
There was a problem hiding this comment.
Using a release candidate for now. @mattersoflight I think we can aim for a 'stable' release after testing and merging these tutorials, so the version that is deployed in production can point to it.
cc @edyoshikun
There was a problem hiding this comment.
I think the tutorial reads well. The huggingface demo does allow inference with VSCyto2D. So the differential value of VCP tutorial is to demonstrate all 3 key models. @ziw-liu let's add them since we also describe them in model card.
| # ?FullyConvolutionalMAE | ||
|
|
||
| # %% | ||
| vs_cyto_2d = FcmaeUNet.load_from_checkpoint( |
There was a problem hiding this comment.
How about we illustrate inference with vs_cyto_3d and VSNeuromast?
There was a problem hiding this comment.
It might be best if we just have and load yml config files for each case.
| # Adjust based on available memory (reduce if seeing OOM errors) | ||
| batch_size=8, | ||
| # Number of workers for data loading | ||
| # Set to 0 for Windows and macOS if running in a notebook, |
There was a problem hiding this comment.
In the past we had set the num_workers=0 given the Windows/Mac issue.
There was a problem hiding this comment.
Google Colab always runs on Linux so I'm using MP to make it faster.
Tutorial notebooks for the Virtual Cell Platform.