Skip to content

Conversation

@marcosolime
Copy link

@marcosolime marcosolime commented Oct 29, 2024

The way you pick valid windows does not seem correct. In your np.random.choice, you are providing as input the length of boolean mask (projs_window_valid_indx.shape[0]), corresponding to valid and invalid windows, but in this way you may pick also False windows.

Instead, you want to return the indices of valid windows only (torch.where). Then, you generate a random permutation of the valid indices and select the first self.window_num.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant