Implemented pre-trained CNN model-architectures to distinguish Dogs from everything else in this world :), followed by identifying their correct breeds.
In addition to this, I also implemented and compared the results from 3 great architectures, namely, AlexNet, ResNet, & VGG and found out which one works the best in most of the cases (including edge cases).

1. Move inside 'uploaded_images' directory

2. Delete(or keep) previous files and upload new images.
   (Note: Image should be Sqaure in shape, i.e. their length and breadth should be similar px.)

3. Open the terminal and run the below script :

   
   $ sh run_models_batch_uploaded.sh
   
   

Try experimenting with variety of images such as :

1. animals other than dog,
2. any random object,
3. or a same picture with different angle would work too !

Results clearly show that VGG shines out.. 🎉🌟

For a more detailed version of the results, check out the respective files for each model in the following format : <model_name>_pet-images.txt & <model_name>_uploaded-images.txt