Clément Pinard

Hi, you can easily log the ground truth with the `tensor2array` function. Here, in a very similar project, I made a little function that logs best and worst results that...

Hi, 1. No, it's juste a vizualisation. If you want to output actual depth, you can this line [here](https://github.com/ClementPinard/SfmLearner-Pytorch/blob/master/run_inference.py#L77). Careful, it will be much heavier as it's uncompressed floats. ```python...

Reported results are trained on Eigen's split. I only now realize that odemetry results should be reported with a training on odom split ! Thanks for that ! I'll come...

You can probably now use tensorflow's leakyrelu since it's supposed to be the same as in lmbspecialops https://www.tensorflow.org/api_docs/python/tf/nn/leaky_relu

I did not write them (I am not the original author) I never downloaded it so I don't really know how to use, I'll see if it's possible to add...

pose estimation and depth are trained at the same time. If you want to train on KITTI odometry 00-08 so that you don't have overfitting on the 09-10 scenes, I...

Hi, indeed you are right ! Thanks for this catch ! If you submit a PR I'd happily merge it.

Yes, the problem is there as well. I will validate this change in your PR :)

Hi, I highly suspect the missing normalization. Depth and pose models have not been trained together in the pretrained network folder, and thus warping is not expected to be perfect,...

Hi, see #125 as to why we don't divide by 255. You results are indeed a bit weird. I'll try to reproduce the results to see if something is wrong,...