search

nickgkan / butd_detr

Code for the ECCV22 paper "Bottom Up Top Down Detection Transformers for Language Grounding in Images and Point Clouds"
Other
74 stars 11 forks source link

965 977 excluded? #39

Closed ucaszyp closed 11 months ago

ucaszyp commented 11 months ago

Thank you for your work.

Could you please explain the meaning of the following code block? Why were scenes 965 and 977 excluded?

if self.split == 'train': annos = [ anno for a, anno in enumerate(annos) if a not in {965, 977} ]

nickgkan commented 11 months ago

Hi, we had some trouble with misaligned/missing annotations so we excluded those two scenes from our training set. Note that we use all scenes of the test set so our results are still comparable to other approaches.