Rene Ranftl

We plan to release the training code. I can't give an exact timeline at the moment, but I hope that we'll be able to do this within one or two...

@Tord-Zhang: We typically train on 4 Quadro 6000 cards that have 24 GB memory each. A complete run to produce the final model takes about 5 days to complete. @angrysword:...

We don't go through all the images in every "epoch". Since the sizes of individual datasets can differ by an order of magnitude, we use a resampling strategy that assembles...

The steps are correct to align the estimates to the SfM construction, but SfM is unable to recover absolute depth too. So the aligned depth maps would have a consistent...

Yes to both questions. A word of caution for evaluating the large model this way: when evaluating the existing large model, which doesn't estimate absolute depth, the numbers are not...

As the results of the model are not perfect a residual error is expected. How much error, will likely vary per image. Here are some works that try to address...

Have a look at the discussions in #36, #37 and #42. In the context of benchmarking you could do what we did in the paper: estimate a scale and shift...

Yes, the output of the auxiliary layer is upsampled to the shape original size using bilinear upsampling. The auxlayer is specified in the inference code. It is applied to "path_2"...

Seems to be similar to #56. You might be running an older version of PyTorch. Can you upgrade to PyTorch 1.7 and try again?

Great! With respect to equirectangular images: we haven't tried, but would be curious to hear what you find out. The extreme distortions encountered in spherical cameras are not present in...