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
I would like to run your code, but I noticed that it only accepts the CLEVR dataset in the .png format.
Unfortunately, I only have the data in .tfrecord format from the official multi-object-dataset.
Should I use the original CLEVR dataset from https://cs.stanford.edu/people/jcjohns/clevr/ (though it does not provide mask labels) or should I manually convert the multi-object-dataset to the .png format?
It would also be extremely helpful for newcomers in this field like me if you could specify the source of each dataset. 😄
The text was updated successfully, but these errors were encountered:
Hello, thank you for sharing the excellent code!
I would like to run your code, but I noticed that it only accepts the CLEVR dataset in the .png format.
Unfortunately, I only have the data in .tfrecord format from the official multi-object-dataset.
Should I use the original CLEVR dataset from https://cs.stanford.edu/people/jcjohns/clevr/ (though it does not provide mask labels) or should I manually convert the multi-object-dataset to the .png format?
It would also be extremely helpful for newcomers in this field like me if you could specify the source of each dataset. 😄
The text was updated successfully, but these errors were encountered: