CompRx Variational Autoencoder

[StrongTanisha]

Source / repo

https://github.com/StanfordMIMI/CompRx

Model description

Medical images are often acquired at high resolutions (>1 megapixel) in order to capture fine-grained details necessary for diagnosis. CompRx is a benchmark with clinically-relevant evaluation tasks that measure the preservation of fine-grained diagnostic features in compressed medical images.

Dataset

Literature benchmark source

https://openreview.net/forum?id=ZiNFhNFxMf

Literature benchmark performance

Strong Compute result achieved

[VALUE/S]

Basic training config (as applicable)

Nodes: 12 Epochs: [N] Effective batch size: 72 Learning rate: [L] Optimizer: Adam

Logs gist

[URL]

[StrongTanisha]

Roadblock - requirements in the repo were incorrect. Not much instruction on how to run it