Wasserstein Generative Adversarial Network #2660
Conversation
Wasserstein Generative Adversarial Network
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #2660 +/- ##
=======================================
Coverage 83.19% 83.19%
=======================================
Files 819 819
Lines 106748 106761 +13
=======================================
+ Hits 88811 88824 +13
Misses 17937 17937 ☔ View full report in Codecov by Sentry. |
There was a problem hiding this comment.
Awesome!
Glad to see you got your example to work after all our small discussions here and there 😄
I'll take some time to review this later today or tomorrow and test it, do your samples look similar to the pytorch example?
The CI is complaining about some formatting if you have some time to fix that.
Thanks for the PR 🙏
|
Yes, they look similar. Here is an comparison with pytorch, and you can find more images generated by pytorch wgan on https://bytepawn.com/training-a-pytorch-wasserstain-mnist-gan-on-google-colab.html |
There was a problem hiding this comment.
Sorry for the late review 😅 was busy with the preparation for yesterday's release.
The example is good! Instead of commenting for the restructure to follow the other workspace examples I made the changes myself.
The performance could probably be improved on GPU, but at least now we have a GAN example 😄
Wasserstein Generative Adversarial Network imlemented in Burn