This project uses Siamese networks and Classification based approach to identify a whale image. The dataset used for this project is available from Kaggle
- Python3
- Pytorch
- Torchvision
-
Data is to be downloaded in a directory named data which you have to make in the git folder. An example of sample relative image path data/train/d9a83d92.jpg
-
Install all the libraries.
-
Type in terminal
python3 train_test_whale_siamese_run_15.pyIf you have multiple GPUs and you want to use a single GPU, type
CUDA_VISIBLE_DEVICES=1 python3 train_test_whale_siamese_run_15.pyOutput is shown as a ratio. Multiply by 100 to interpret it as a percentage. eg: if val acc is shown 0.1234 it means the accuracy is 12.34%
- Abhinav Kumar, University of Utah
- Surojit Saha, University of Utah