Max Zhaoshuo Li 李赵硕

Hi @aakash26, thanks for reaching out. As mentioned in Q&A item 3, there are two types of occlusion, 1) at left image border, 2) at left border of objects. -...

Hi @aakash26, Is your dataset similar to KITTI? - If so, you can use the KITTI pretrained weights. - If not, you should use the Scene Flow pretrained weights. This...

Did you use one of these from https://pytorch.org/get-started/previous-versions/? ``` # CUDA 10.2 conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=10.2 -c pytorch # CUDA 11.1 conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=11.1 -c...

Hi @EhrazImam I am sorry for the problems you have got. But this is a more appropriate question to ask in the PyTorch forum. Can you post your question here:...

depth = f * b / disp, where f is focal length and b is baseline

Hi @ynjiun Thanks for reaching out. I assumed you are using the KITTI pretrained model? The KITTI pretrained model is severely overfitted to the small dataset. For better generalization, please...

Hi @Miaowei-HNU The occlusion is computed from the following steps - The provided depth map are backprojected to both left and right cameras using the preprocessing scripts [here](https://github.com/mli0603/MICCAI_challenge_preprocess) - The...

hello @btwbtm Given a sequence of W, the total **signed** relative positions will be 2W-1. For example, if you have W=3, the relative positions will be `-2, -1, 0, 1,...

Hi @GREW-Benchmark The code currently doesn't support batch size larger than 1 because of the random cropping. You can comment it out for larger batch size.

Hi @GREW-Benchmark Thanks for the update. There were two reasons the code doesn't support larger batch size: - I only have one gpu and cannot perform larger experiments - Compared...