Grégoire Passault

For some application, I would also be interrested in having the values predictions available, so I am adding them as well, some questions: * Are the Q/V networks decent estimations...

Also, is there a consistent pattern across all algorithms to access values functions just like `predict` for the action ?

Another question/note: If we use the `normalize` hyperparameter while training, I guess we will need to use the normalization during the inference, right? That would imply getting some data from...

> @Gregwar I could successfully test it =) That is nice! > I had to do some tweaks to use it with conda env: Yes you are right this should...

Hello, Sorry for the lag, we are currently working on our humanoid robots for RoboCup, we integrate DRL algorithms in the robots for the first year. We spent some time...

Hello Ryan, Thanks for your PRs! I will likely not be able to review them anytime soon since I am currently preparing for RoboCup 2022 that will be held next...

Hello, I am not sure to understand what your problem is exactly Can you explain with more details? Like what you expect compared to what you actually get ? Joints...

Hello, You are right This should be fixed in 0.3.5, can you confirm?

Hello, I agree that it might be easier, but it would require deeper rework of the logic Actually, some current constraints can be lightened by refactoring/rewriting onshape-to-robot, hopefuly I will...

Just for the detail; currently the problem is also that non-dof relation (like two parts being fastened together) don't work on top-level assembly (that could also be improved)