CSCE 496/896 (Spring 2018) Hack-A-Thon 4: Convolutional Layers 1


Assigned Monday, February 5
Due Thursday, February 8 Saturday, February 10 at 11:59 p.m.


First, download the file hackathon_4.ipynb. Then, go to crane.unl.edu and log in via two-factor authentication. You then need to select "Local Jupyter Notebook" and click "Upload" to upload hackathon_4.ipynb

The MNIST data and submission scripts are on crane in the folder

/work/cse496dl/shared/hackathon/04
This folder contains the script run_py.sh for submitting optimization jobs to crane. Also in this folder is the subfolder mnist, which has four numpy files: mnist_test_images.npy, mnist_train_images.npy, mnist_test_labels.npy, and mnist_train_labels.npy. (These files are also zipped up in mnist.zip, but you should use the ones already on crane for this assignment.) The Jupyter Notebook for this hack-a-thon contains instructions for reading data from these files and partitioning the test files into validation and optimization sets.

Now you should be able to step through the notebook. When done, select "File" → "Close and Halt". Then check the box next to your notebook and click "Download". Then go to handin and submit your ipynb file by 11:59pm Thursday.


Return to the CSCE 496/896 (Spring 2018) Home Page

Last modified 06 February 2018; please report problems to sscott.