This assignment consists of the following:
- Image_Captioning, where you'll examine a classic image captioning dataset. You'll also read one of the classic image captioning papers and answer some questions about their architecture. Finally, you'll experiment with CLIP embeddings to categorize the content of images.
- answers file where you'll put all your answers (except D5)
This assignment does not require you to use a Colab notebook. You can run it all on your GCP instance and then use the submit.sh script to submit your work. As a result you do not need to run scp to copy notebooks this time.
As with Assignment 3, please submit by running the submit.sh script, only with -a 4 (since this is assignment 4).
./assignment/submit.sh -u your-github-username -a 4
It is your responsibility to check that your work has made it to your GitHub repository in the a4-submit branch. As always, a small number of points will be added for submitting correctly. We will give each person who correctly submits their assignment one bonus point on this homework assignment.