Hello, thank you for open-sourcing this excellent work on Distributional RL Navigation.
I am currently trying to reproduce the project, but I encountered several dependency and compatibility issues related to Gym and Stable-Baselines3 versions.
Specifically, I encountered:
- Gym API compatibility issues (
reset() interface mismatch)
ObsEncoderPolicy unknownsb3_contrib dependency issues- Potential incompatibility between SB3 2.x and the custom policy registration
Could you please provide the exact runtime environment used in the paper/code release?
It would be extremely helpful if you could share:
- Python version
- PyTorch version
- Gym version
- Stable-Baselines3 version
- sb3-contrib version
- NumPy version
- CUDA version (if applicable)
- Full
requirements.txt or environment.yml
I would greatly appreciate your help. Thank you again for releasing this valuable work.
Best regards.
Hello, thank you for open-sourcing this excellent work on Distributional RL Navigation.
I am currently trying to reproduce the project, but I encountered several dependency and compatibility issues related to Gym and Stable-Baselines3 versions.
Specifically, I encountered:
reset()interface mismatch)ObsEncoderPolicy unknownsb3_contribdependency issuesCould you please provide the exact runtime environment used in the paper/code release?
It would be extremely helpful if you could share:
requirements.txtorenvironment.ymlI would greatly appreciate your help. Thank you again for releasing this valuable work.
Best regards.