Dog Breed classifier built on bottleneck features extraction using pretrained model (Densenet, resnet50v2, nasnet) and then stacked together to get better output and then trained on custom DNN and then deployed as website using streamlit and then deployed on heroku server
weblink :https://doggfind.herokuapp.com/
Dataset Used :https://www.kaggle.com/c/dog-breed-identification
Notebook Link : https://www.kaggle.com/preatcher/dog-breed-classifer-using-pretrained-model