self-supervised-depth-completion icon indicating copy to clipboard operation
self-supervised-depth-completion copied to clipboard

inference

Open hello7623 opened this issue 5 years ago • 2 comments

What should I do if I only want to use your code for a few images?

hello7623 avatar May 14 '20 14:05 hello7623

@fangchangma can you please provide us the instructions that we can follow to make inferences using your pretrained models?

@hello7623 or did you already figure that out? If yes, could you please tell how did you do that?

Rajat-Mehta avatar Jun 09 '20 09:06 Rajat-Mehta

My findings, while looking around in the code and trying around:

  • there is no plain inference method.
  • You cannot use the pretrained model on only images. As far as I understand, the pretrained model uses the input 'd' and 'rgb', which means both sparse depth and rgb image data. Therefore you won't get around training the model yourself if you have no depth data.

kfina91 avatar May 17 '21 13:05 kfina91