learnopencv/Medical_Multi-label_Classification_with_PyTorch_&_Lightning

Medical Multi-label Classification with PyTorch & Lightning

This repository contains the files related to the LearnOpenCV blog post: Medical Multi-label Classification with PyTorch & Lightning

This project was created using the subset of Human Protein Atlas Image Classification challenge dataset.

The subset dataset includes the top 10 classes of the above dataset and was used as an in-class competition here: Zero to GANs - Human Protein Classification.

As the competition is ended, the subset dataset can be downloaded from over here:

1. Subset - 512x512 size
2. Subset - 384x384 size (used for the project)

We used EfficientNetV2-Small model from Torchvision for training and deployed the app using Gradio onto HuggingFace spaces.

You can use the demo app here --> Medical Multi-Label Image Classification Gradio App

You can also check/access and download the following files:

1. app.py python file.
2. Trained lightning checkpoint file.
3. requirements.txt for the gradio app deployment on HuggingFace

From here --> Gradio App Files

You can run it on your local system or on Colab.

---

---

AI Courses by OpenCV

Want to become an expert in AI? AI Courses by OpenCV is a great place to start.