Open Sylva-Lin opened 3 weeks ago
The test code iterates multiple times through partial results, accumulating them to obtain the final prediction. Is this approach similar to the Test-Time Augmentation used in other methods?
Yeah, it is a kind of TTA. I more detailed explanation is here: https://github.com/Pointcept/Pointcept?tab=readme-ov-file#testing
The test code iterates multiple times through partial results, accumulating them to obtain the final prediction. Is this approach similar to the Test-Time Augmentation used in other methods?