Replies: 2 comments
-
Hi, you can customize a However, visualizing the training data during training will consume a lot of time, we suggest using this script like this to check the augmentation result. |
Beta Was this translation helpful? Give feedback.
0 replies
-
Thanks for your quick response and the references. You're right about that, probably better to have an offline script iterates through data loader (so custom sampling methods can be checked too). |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
What is the feature?
Are there any existing functionality to visualise (a subset of) training batch? This will be useful if one is using strong/special augmentation or sampling methods, to allow users to know that the training input images are as expected. Should be straightforward to implement as a
before_train_iter
hook?Any other context?
Thank you!
Beta Was this translation helpful? Give feedback.
All reactions