Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- (Related XKCD Comic from 2014: https://www.explainxkcd.com/wiki/index.php/1425:_Tasks )
- We will identify whether an image belongs to a Cat or a Dog. Unlike the Credit Card Default assignment, this data set is too large to process on the virtual machine, and you must use the GPU accelerated server.
- You will be submitting your analysis to https://www.kaggle.com/c/dogs-vs-cats-redux-kernels-edition
- Part 1: Resize All Images to 32x32 pixels. (You can generalize your code to also work on 64x64, which may have higher accuracy)
- Part 2: Train your deep convolutional neural network (model.fit)
- Please use the MNIST and chars74k examples from class as a starting point.
- Part 3: Submit your predictions (model.predict_classes) to Kaggle
- I expect this assignment may take more than one week, since model run-times can sometimes take several hours. Please start with small model hyperparameters to test your python code before expanding the depth/width of your network.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement