Check out this awesome Streamlit app I built
https://share.streamlit.io/gunjandhanuka/pokedex_classifier/app.py
GitHub Link with Documentation: GitHub - GunjanDhanuka/PokeDex_Classifier: A web app that classifies an image into one of 150 Pokemon using Transfer Learning and Streamlit Framework
- This tool will help you identify the Pokemon you encounter in your way in the Kanto Region (Generation 1).
The model uses Transfer Learning from the DenseNet201 model to classify the images into 150 different classes. First the images are resized according to the model input and then using a Softmax layer at the end, we compute the probability of the image to be one of 150 Pokemon. The model might have some difficulty in differentiating between evolved forms of a Pokemon, for example Pidgeotto and Pidgeot!. However you are free to try your own quirky images as well