-
Notifications
You must be signed in to change notification settings - Fork 417
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
added Generating fonts with adversarial networks example #94
base: master
Are you sure you want to change the base?
Conversation
Neat stuff! Have you played with doing a conditional version? Do you know the license of the notmnist dataset? It would be best to rehost it in the Recipes S3 bucket if allowed, and maybe a dump of the trained model weights as well. |
How can we rehost it in the s3 bucket? No, I haven't tried the conditional version yet. I will upload pre trained weights and update the notebook as soon as i get to computer. |
…xamples/Generating fonts with adversarial networks/Generating fonts with adversarial networks.ipynb
As for conditional version, i did a tiny rework of @TIXFeniks 's notebook to make it 'conditional' and it does work: |
Could you please repeat how to upload to AWS? |
That's cool. I think I'd prefer the conditional version if we are only keeping one of them. (Reworking the notebook so the conditioning is only added for an additional experiment in the end would be nice, but seems to be cumbersome.) We'd need some explanatory words, though. Before "Before we dive into training them", you could add some notes on (a) that you add conditional variables to make it easier to learn and to allow conditional sampling and (b) that G(z) and D(x) will thus change to G(z, y) and D(x, y). And the beginning of "Generative models 101" has a broken image.
Only us maintainers can do so. We need to know which files to upload and then we can send you the link to include in your notebook. Would you like to change this PR to include the conditional version? Or shall we add both? (I wouldn't mind having both |
Fixed what you asked. By the way, we've lost our GPU for a while. Could you train the networks yourself and upload pretrained weights? |
added Generating fonts with adversarial networks example