Example RL algorithms or other repos utilizing RLBench?

[ahundt]

I'm looking into using RLBench, thanks for creating it! Are any baseline algorithms already integrated in this repository or another one?

[stepjam]

Hi, RLBench is a set of environemnt, and so we wont ever put any baseline algorithms in this codebase. If you are looking for baseline RL algortihms, then I suggest one of:

• OpenAI Baselines
• OpenAI Spining Up
• RLLIB/RAY

You can then use the algorithms there to train on these tasks :smile: If you are familiar with OpenAI Gym, then RLBench has recently become gym compatible. See here.

[ahundt]

So you’re saying there is no other RL repository which already has this one integrated? Correct me if I’m mistaken, but if there are no known, working, and integrated clients I feel like there will be a lot of debugging to do in this project and it may not be ready for use unless one is willing to go through debugging the whole API themselves. Im not trying to criticize, rather I’m asking to understand the scope of work if I were to use this project seriously. I have a paper & project which uses the vrep api and I’m figuring out the possibility of integrating that with this work https://github.com/jhu-lcsr/good_robot.

[id9502]

I'm looking into using RLBench, thanks for creating it! Are any baseline algorithms already integrated in this repository or another one?

Hi, I am working on adding some RL/IL algorithms upon this great environment for long term tasks. I am now making some training on Mujoco envs and will soon start my experiments on this env. I will keep updating related results on my repo once I have finished my tests on: id9502/RLFrame.git