yuliangguo/OmniFusion

Issues

• Pretrained checkpoint

  #13 opened a year ago by kimkj38
  1

• Could you provide the pretrained model for Matterport3d?

  #11 opened 2 years ago by EchoTHChen
  2

• performance on Stanford2D3D is worse than reported

  #7 opened 2 years ago by apple2373
  4

• The evaluation results depend on batch size

  #10 opened 2 years ago by apple2373
  1

• It should not be drop_last=True for testing.

  #9 opened 2 years ago by apple2373
  1

• is it a common practice to ignore errors < 0.1?

  #8 opened 2 years ago by apple2373
  2

• Size mismtach for pre-trained model

  #3 opened 2 years ago by babar41
  7

• dataset_sem is missing

  #6 opened 2 years ago by apple2373
  0

• Inference

  #2 opened 3 years ago by pablodawson
  8

• Could you provide the pretrained model for Matterport3d?

  #5 opened 2 years ago by EchoTHChen
  2

• Reason for using Conv3d instead of Conv2d

  #4 opened 2 years ago by MertCokelek
  1

• Datasets

  #1 opened 3 years ago by libetter0913
  0