facebookresearch / consistent_depth

We estimate dense, flicker-free, geometrically consistent depth from monocular video, for example hand-held cell phone video.
MIT License
1.61k stars 236 forks source link

Colab does not generate any /result folder #49

Open dpredie opened 3 years ago

dpredie commented 3 years ago

it completes without error, saying results are available in /content/consistent-depth/result/ but no such folder generated

dpredie commented 3 years ago

i find that it is not able to do import wget on the final block

Ebernn commented 3 years ago

same issue, that's probably because of ./scripts/install.sh which doesn't do its job (unable to install dependencies correclty) 😒

Ebernn commented 3 years ago

so this issue might be a consequence of #42

Z-XO commented 2 years ago

Has anyone found a solution to this? The version of torch is 1.4 and cuda 10 , as required in the requirements text. I am unable to reproduce similar results to that presented in the paper. Even while using a similar input, slow panning video on an object that does not move that much

EntityinArray commented 2 years ago

Experiencing the same issue. Fix the colab notebook, please