This project involved trying to use CLIP (OpenAI) for the detection of different lymphoma types
Next goals:
- Try and make an HDF5 backend for the data - similar to what was done in the dense net blogpost HDF5 backend is not properly optimized - need to go over data preprocessing (CODE from paper is not optimal)
- Figure out why the classification is so bad (?)