HW6 (Bonus). Image Classification, Data Augmentation, Pre-Trained Models (28 May, 23.59)
EX1 (5p) : This exercise is to build a classifier using the Cats v Dogs dataset of 25k images! sed You have to train a convolutional neural network for the classification.
You will see the CNN overfit very quickly, despite great results with the training set.
Complete the notebook and upload your results.
https://colab.research.google.com/drive/1SsdYGrvglje8bhPc0GuFcBSJs9Jt0mor#scrollTo=6f375c5d
EX2 (5p) : This excercise is an extension to previous EX1. A data augmentation is build to tame the overfitting. You need fill the data augmentation and run the code.
Complete the notebook and upload your results.
https://colab.research.google.com/drive/1Crn2BGalP7obGX5jAApUQqJRrjTsKwAg#scrollTo=35525d12
EX3 (15p) : This excercise is to build classifier with a pre-trained CNN.
Complete the notebook and upload your results.
https://colab.research.google.com/drive/1x9T_89sn1U_SBRPX61bAI9tSjQoxoSXt#scrollTo=bZTBsWIbG0Ju