search

TempleRAIL / drl_vo_nav

[T-RO 2023] DRL-VO: Learning to Navigate Through Crowded Dynamic Scenes Using Velocity Obstacles
https://doi.org/10.1109/TRO.2023.3257549
GNU General Public License v3.0
149 stars 11 forks source link

Versions and Hardware specs #1

Closed lucy2329 closed 1 year ago

lucy2329 commented 1 year ago

Hey, I was trying to run the baseline model that you have mentioned on the paper. I see that you're using a DGX-1, could you provide some information on how long it took. Also, stable-baselines3 has a pre req of torch > 1.11 as shown here. Is there a requirements.txt I can go off of?

zzuxzt commented 1 year ago

Thank you for your interest in our work. The ReadMe introduction is still under construction. I will update the code with more detailed information, requirements file, and add a container environment with required packages this weekeed. However, here is some information you may need:

  1. I only used 1 GPU in DGX-1 and it took about 1 week for training.

  2. The new vesion of stable-baseline3 needs torch > 1.11, but the old one does not require it. The detailed version information of learning packages is as follows, and you can add them to a requirements.txt to install:
    gym==0.18.0 pandas==1.2.1 stable-baselines3==1.1.0 torch==1.7.1

Or you can manually install them by the following commands: pip install torch==1.7.1+cu110 -f https://download.pytorch.org/whl/torch_stable.html pip install gym==0.18.0 pandas==1.2.1 pip install stable-baselines3==1.1.0