This repo includes code (written in Python) for Caltech-UCSD Birds-200-2011 dataset classification. I have used PyTorch Library for CNN's. You can download the dataset here http://www.vision.caltech.edu/visipedia-data/CUB-200-2011/CUB_200_2011.tgz
This Project is done using Google Colaboratory (Online free GPU support provided by Google(Thanks to Google for this)). This dataset contains birds images of 200 categories. Training images are 5994 & test images are 5794. I have achieved top1 accuracy 77.32% and top5 accuracy 94.30% on test set. This project is done using Transfer learning techniques with a ResNet18(pretrained on imagenet). You can get more details in the notebook file.
The result probabilities looks like: ![] (https://github.com/Muhammad-MujtabaSaeed/Caltech-Birds-Classification/blob/master/bird%20result.PNG)