CVLAB-Unibo/Learning2AdaptForStereo

Questions about dataset and devices.

Pengxiang-Li opened this issue · 1 comments

Pengxiang-Li commented

hello! thanks for the great work!
When I tried to run your code on my pc, I cannot run it directly and I have 3 questions about your work.

  1. How do you organize your ' adaptation_list.csv '?
  2. The order of training on Sceneflow, kitti and other datasets.
  3. When you training the models, how many GPUs did you use? And their video memory?

thanks again!

AlessioTonioni commented

Hi,

Regarding 1 what do you mean exactly? The details on the dataset preparation are reported here: https://github.com/CVLAB-Unibo/Learning2AdaptForStereo#data-preparation. It boils down to having one CSV per stereo sequence with consecutive frames in conseutives lines.

Regarding 2, the order of the datasets is reported on the paper, it should be SceneFlow (without meta learning) then either Kitti, Carla or synthia (with meta learning)

Regarding 3, all the result in the paper were achieved with a single GPU, I'm not exactly sure on the model but something equivalent to a Titan X with 12 Gb of VRAM