This repo covers Cola_SuperBin group's solution to the second task(image quality assessment) of DRAC 2022 Diabetic Retinopathy Analysis Challenge(DRAC).

We download the dataset from DRAC 2022.

The dataset contains a training set consisting of 665 images and test set consisting of 438 images.

1. Clone our repository:

   git clone https://github.com/Otsuts/DRAC22.git
   cd DRAC22

2. To run our code, you can use the command:

   python main.py --model <model_name> --batch_size <batch_size> --num_epochs <training epochs> --lr <learning rate> --weight_decay <weight decay> --save_model <save model> --device <which device> --transform <augmentation> --k_fold <ensemble>

Our team Colar_Super_Bin ranks 1st on the leaderboard!