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,
In the data.py file, we can see the names of 919 classes provided. However, the indices in ImageNetS_categories_im919.txt do not correspond to the classes. How should I go about resolving this issue?
Looking forward to your response.
The text was updated successfully, but these errors were encountered:
Sorry, not sure I understand your question. The file is the matching between imagenet-s class id and imagenet tag id. Do you mean there is something wrong in the ImageNetS_categories_im919.txt?
My understanding is that ImageNetS_categories_im919.txt corresponds to each of the classes in classes_919 in turn, but I've found that this is not the case.
So is there any way I can map each class in classes_919 to the serial number of train_semi?
I found that both classes_50 and classes_300 correspond to ImageNetS_categories_im50.txt and ImageNetS_categories_im300.txt.
Hi,
In the data.py file, we can see the names of 919 classes provided. However, the indices in ImageNetS_categories_im919.txt do not correspond to the classes. How should I go about resolving this issue?
Looking forward to your response.
The text was updated successfully, but these errors were encountered: