Hello! There are lots of errors in the code, please recorrect the README.txt, so that can run the result

montrealrobotics / one4all

An end-to-end fully parametric method for image-goal navigation that leverages self-supervised and manifold learning to replace the topological graph with a geodesic regressor.

MIT License

8 stars 1 forks source link

Hello! There are lots of errors in the code, please recorrect the README.txt, so that can run the result #3

Open xiaofeifei-1 opened 4 months ago

Jaeyoung-Lim commented 2 weeks ago

@MikeS96 @sachaMorin I have tried following the readme to reproduce the results from the paper, but it seems that the readme has a few issues when trying running the code

sachaMorin commented 2 weeks ago

Hello, can you share the errors? Are you trying to train components or only deploying the checkpoints?

Jaeyoung-Lim commented 2 weeks ago

@sachaMorin I am trying to reproduce the results in the paper, mainly the point navigation.

I would (hopefully) like to try and use this approach on my own dataset for a robotics application.

I have posted my initial problems in https://github.com/montrealrobotics/one4all/issues/4
I still haven't got to run the code, but some minor fixes here: https://github.com/montrealrobotics/one4all/pull/5 I could try to fix the readme as I run through the instructions.

MikeS96 commented 2 weeks ago

Thanks for pointing this out. We made some suggestions in #4. Please do let us know if it helps.