This homework is due on October 5, 2021. You will learn to train multi-label image classification models using the Pytorch framework. We will classify images from the FashionMNIST and PASCAL 2007 datasets into the object(s) present in the image.
Instructions:
- Starter code and full assignment can be found in this repo
- Please submit your solutions using the class Gradescope.
- Your hand-in should consist of two parts: 1) a single PDF report (print-out from Jupyter is fine) and 2) all your code in a compressed tar file. Please name the report "AndrewID.pdf" and the code "AndrewID.tar".
- You should follow the code structure we defined in the steps of this assignment, but small deviation is allowed.
- It is okay to ask Google for help, but it is not okay to plagarize. Feel free to search “how to add cross entropy loss in pytorch,” but do not steal code from online repos. We are familiar with online tutorial code just so you know....
- For the programming assignments, students will be allowed a total of five late days per semester; each additional late day will incur a 10% penalty. The code should be easy to run by TAs.
- Start early.... please!
- Required: Have fun 😜
Acknowledgements: This assignment was adapted from previous homework written by Yufei Ye and Sudeep Dasari, who in turn took code from the PyTorch tutorial. Many thanks to the original authors!