- Modify the generator of DCGAN to exactly match the paper—are you able to faithfully recreate their results? Why or why not?
- Modify the download script and download a different LSUN dataset. Did you need to modify your code to get results? What parameters were important?