- pytorch = 1.8
- cuda 10.2
- torchvision
- matplotlib
Uncompress the folder in any location and run the following command
python test_geo.py --train PATH_TO_TRAINIG_DATA
The training data folder must have 2 folders called set_test and set_train
For testing you can run the following command
python test_geo.py --test PATH_TO_TEST_IMAGE --k 10
where k is the number of top k images similar to the testing one
https://github.com/matthias-wright/cifar10-resnet https://github.com/lambdal/cifar10-fast https://medium.com/swlh/natural-image-classification-using-resnet9-model-6f9dc924cd6d https://github.com/digantamisra98/Mish