You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, that is a great work, but I have a problem. Such as in "kmeans_clustering_search_human.ipynb", you use the trained model "output_path = 'human_ada.pth". My problem is when you use K-means, why it can directly cluster two eyes or two ears together. So when you trained the stylegan model, do you need the classes label?
The text was updated successfully, but these errors were encountered:
@warmspringwinds Do you have any advice to make left and right ears to be in separate clusters? My first guess is using some kind of heuristic, but can we make it from the clustering step? I tried to increase K but it doesn't
Hi, that is a great work, but I have a problem. Such as in "kmeans_clustering_search_human.ipynb", you use the trained model "output_path = 'human_ada.pth". My problem is when you use K-means, why it can directly cluster two eyes or two ears together. So when you trained the stylegan model, do you need the classes label?
The text was updated successfully, but these errors were encountered: