Project 4 -- Image Generation
Due Date: 04/30/2026 23:59
Download
[Jupyter Notebook]
Colab:
[Link]
Late Policy
- Late submissions will be allowed up to two days after the original deadline, but may be abridged (e.g. due to the end of the semester). Late submissions will be penalized 20%.
- See Gradescope for specific due dates and late submission deadlines.
- Regrade requests must be made via GradeScope within a week of grades being released.
Project 4 – Image Generation
Your task for this project is to build a generative model for cute animal pictures based on a provided data set. You may use any of the generative modeling types discussed in class so far (GANs, VAEs, and diffusion models), but you are encouraged to try and implement diffusion models.
Because the data set you are using is already copied to SCC, you may want to copy the notebook to your SCC directory and run it there. The easiest way to do that is to run the following commands:
cd /projectnb/dl4ds/students/YOUR-USERNAME
mkdir project4
cd project4
wget https://raw.githubusercontent.com/DL4DS/sp2026/refs/heads/main/static_files/assignments/project4.ipynb
Do not copy the data set to your SCC directory!!
Then you can create a new interactive Jupyter notebook session from the SCC OnDemand dashboard with the following suggested settings:
- List of modules to load:
miniconda academic-ml - Pre-launch command:
conda activate spring-2026-pyt - Interface:
lab - Working directory:
/projectnb/dl4ds/students/YOUR-USERNAME - Number of CPUs:
4 - Number of GPUs:
1 - GPU compute capability:
7.0 - Project:
dl4ds
Note: You can try a lower GPU compute capability than 7.0 and in fact you will get a GPU compute node more quickly, but it seems that the
academic-mlmodule requires GPUs of at least version 7.0.
Turn in on Gradescope.
Have fun!