An End to End deep learning project for identifying 250+ bird species. React is used for frontend and flask as backend PyTorch is used for training the deep-learning model
Dataset is taken from kaggle, which has 275 species dataset.
ResNet-50 model is trained on Google colab pro which is implemented in PyTorch deep learning framework.
Achieved 99.9% accuracy on trainig set and 97.1% on test set, there is a chance of imrovement by adding data augmentation. It is observed that model struggle to identifies birds with front profiles.