search

isaac-sim / IsaacGymEnvs

Isaac Gym Reinforcement Learning Environments
Other
1.89k stars 409 forks source link

Documentation should be updated #17

Closed noshaba closed 2 years ago

noshaba commented 2 years ago

Hi,

The documentation, e.g. in https://github.com/NVIDIA-Omniverse/IsaacGymEnvs/blob/main/docs/framework.md should probably be updated.

gavrielstate commented 2 years ago

Can you be more specific about what you think needs to be updated?

noshaba commented 2 years ago

For example it states that the init method only takes 3 arguments - the config file, sim_device and headless. Yet, in the examples, they also include graphics_device_id. Also the parameters required for the config files and what they mean and which options you have for them would be nice to have written somewhere in a more formal and organized way, like here https://github.com/Denys88/rl_games. I guess the train config files use the same parameters but something like this also needs to be done for the task config files, and at least a reference saying that one can find the definitions of the train config files in that repo would be nice.

noshaba commented 2 years ago

So at least a minimum working hello world example, similar to the Hello World / Robot examples in Isaac Sim.

noshaba commented 2 years ago

Another improvement for the docs could be to add that jit functions should all have unique names. It doesn't work if the same function name for the reward and observation function are used in different tasks but the error messages you get don't indicate that. Only says that the task X could not be imported.