Closed sayakpaul closed 1 year ago
@fchollet did you have the chance to consider it? If not I can definitely understand you might have been swamped :)
What's the TL;DR of what the code does?
It shows how to use a vanilla DCGAN for image classification in a semi-supervised way. The benefit we get here is this technique can really shine when we don't have access to a lot of labeled data. It's, in fact, a minimal implementation of this paper.
I think it's somewhat too experimental and not practical enough to be featured as a code example. But a state-of-the-art semi-supervised classification example would be great I think.
When proposing this idea, my understanding was that a lot of people would be interested in the workflow that uses GANs for semi-supervised learning problems. I will drop it if it does not reflect that thought.
But a state-of-the-art semi-supervised classification example would be great I think.
Any recommendations?
Hi @sayakpaul , Just checking if this is issue is still valid. If not, could you please close the issue. Thanks!
This issue is stale because it has been open for 14 days with no activity. It will be closed if no further activity occurs. Thank you.
This issue was closed because it has been inactive for 28 days. Please reopen if you'd like to work on this further.
Now that the fields of semi-supervised and self-supervised learning are becoming more and more important, I wish to cover an example showing how to train GANs for semi-supervised image classification (can be referred to as weakly supervised learning too).
I have a Colab Notebook that might be useful. If this proposal is accepted I will, of course, adhere to the Keras blog format. I used
tf.GradientTape
(did not overridetrain_step
) as I found it was actually more readable that way.Note: In order to show the full potential of using GANs in this context, one might need to train longer than shown here.