ndb796 / StyleCLIP-Tutorial

51 stars 12 forks source link

training question and sample dataset #4

Open jimb2834 opened 6 months ago

jimb2834 commented 6 months ago

Hello @ndb796

I like the examples you have shown.

I would like to train a new model myself with images I have classified as "man" and "women" with clip and test this.

How many images and "captions" do you think I need just to test the concept? How many did you use for your examples here?

Are you able to share your Training Set and we can re-produce the same results?

Thank you for your time