Data science project using Deep learning for image classification.
Built a classifier to classify an image into 6 categories ->(buildings, sea, forest, mountain, glacier, street). The classifier is built using a Convolution neural network. The architecture used is a modified version of AlexNet.
I got an accuracy of 86% on the test set after iterating and tuning on different architectures.
This Data contains around 25k images distributed under 6 categories.
{'buildings' -> 0, 'forest' -> 1, 'glacier' -> 2, 'mountain' -> 3, 'sea' -> 4, 'street' -> 5 }
This model was trained on a Google Colab GPU runtime which took around 6 seconds per epoch. I recommend using this config as training this on an average computer can take around 10-15 minutes per epoch.