| | .. _tutorial-run-rl-training: |
| |
|
| | Training with an RL Agent |
| | ========================= |
| |
|
| | .. currentmodule:: isaaclab |
| |
|
| | In the previous tutorials, we covered how to define an RL task environment, register |
| | it into the ``gym`` registry, and interact with it using a random agent. We now move |
| | on to the next step: training an RL agent to solve the task. |
| |
|
| | Although the :class:`envs.ManagerBasedRLEnv` conforms to the :class:`gymnasium.Env` interface, |
| | it is not exactly a ``gym`` environment. The input and outputs of the environment are |
| | not numpy arrays, but rather based on torch tensors with the first dimension being the |
| | number of environment instances. |
| |
|
| | Additionally, most RL libraries expect their own variation of an environment interface. |
| | For example, `Stable-Baselines3`_ expects the environment to conform to its |
| | `VecEnv API`_ which expects a list of numpy arrays instead of a single tensor. Similarly, |
| | `RSL-RL`_, `RL-Games`_ and `SKRL`_ expect a different interface. Since there is no one-size-fits-all |
| | solution, we do not base the :class:`envs.ManagerBasedRLEnv` on any particular learning library. |
| | Instead, we implement wrappers to convert the environment into the expected interface. |
| | These are specified in the :mod:`isaaclab_rl` module. |
| |
|
| | In this tutorial, we will use `Stable-Baselines3`_ to train an RL agent to solve the |
| | cartpole balancing task. |
| |
|
| | .. caution:: |
| |
|
| | Wrapping the environment with the respective learning framework's wrapper should happen in the end, |
| | i.e. after all other wrappers have been applied. This is because the learning framework's wrapper |
| | modifies the interpretation of environment's APIs which may no longer be compatible with :class:`gymnasium.Env`. |
| |
|
| | The Code |
| | |
| |
|
| | For this tutorial, we use the training script from `Stable-Baselines3`_ workflow in the |
| | ``scripts/reinforcement_learning/sb3`` directory. |
| |
|
| | .. dropdown:: Code for train.py |
| | :icon: code |
| |
|
| | .. literalinclude:: ../../../../scripts/reinforcement_learning/sb3/train.py |
| | :language: python |
| | :emphasize-lines: 57, 66, 68-70, 81, 90-98, 100, 105-113, 115-116, 121-126, 133-136 |
| | :linenos: |
| |
|
| | The Code Explained |
| | |
| |
|
| | .. currentmodule:: isaaclab_rl.utils |
| |
|
| | Most of the code above is boilerplate code to create logging directories, saving the parsed configurations, |
| | and setting up different Stable-Baselines3 components. For this tutorial, the important part is creating |
| | the environment and wrapping it with the Stable-Baselines3 wrapper. |
| |
|
| | There are three wrappers used in the code above: |
| |
|
| | 1. :class:`gymnasium.wrappers.RecordVideo`: This wrapper records a video of the environment |
| | and saves it to the specified directory. This is useful for visualizing the agent's behavior |
| | during training. |
| | 2. :class:`wrappers.sb3.Sb3VecEnvWrapper`: This wrapper converts the environment |
| | into a Stable-Baselines3 compatible environment. |
| | 3. `stable_baselines3.common.vec_env.VecNormalize`_: This wrapper normalizes the |
| | environment's observations and rewards. |
| |
|
| | Each of these wrappers wrap around the previous wrapper by following ``env = wrapper(env, *args, **kwargs)`` |
| | repeatedly. The final environment is then used to train the agent. For more information on how these |
| | wrappers work, please refer to the :ref:`how-to-env-wrappers` documentation. |
| |
|
| | The Code Execution |
| | |
| |
|
| | We train a PPO agent from Stable-Baselines3 to solve the cartpole balancing task. |
| |
|
| | Training the agent |
| | ~~~~~~~~~~~~~~~~~~ |
| |
|
| | There are three main ways to train the agent. Each of them has their own advantages and disadvantages. |
| | It is up to you to decide which one you prefer based on your use case. |
| |
|
| | Headless execution |
| | """""""""""""""""" |
| |
|
| | If the `` |
| | when training on a remote server or when you do not want to see the simulation. Typically, it speeds |
| | up the training process since only physics simulation step is performed. |
| |
|
| | .. code-block:: bash |
| |
|
| | ./isaaclab.sh -p scripts/reinforcement_learning/sb3/train.py |
| |
|
| |
|
| | Headless execution with off-screen render |
| | """"""""""""""""""""""""""""""""""""""""" |
| | |
| | Since the above command does not render the simulation, it is not possible to visualize the agent's |
| | behavior during training. To visualize the agent's behavior, we pass the ``--enable_cameras`` which |
| | enables off-screen rendering. Additionally, we pass the flag ``--video`` which records a video of the |
| | agent's behavior during training. |
| | |
| | .. code-block:: bash |
| | |
| | ./isaaclab.sh -p scripts/reinforcement_learning/sb3/train.py --task Isaac-Cartpole-v0 --num_envs 64 --headless --video |
| | |
| | The videos are saved to the ``logs/sb3/Isaac-Cartpole-v0/<run-dir>/videos/train`` directory. You can open these videos |
| | using any video player. |
| | |
| | Interactive execution |
| | """"""""""""""""""""" |
| |
|
| | .. currentmodule:: isaaclab |
| |
|
| | While the above two methods are useful for training the agent, they don't allow you to interact with the |
| | simulation to see what is happening. In this case, you can ignore the `` |
| | training script as follows: |
| |
|
| | .. code-block:: bash |
| |
|
| | ./isaaclab.sh -p scripts/reinforcement_learning/sb3/train.py |
| |
|
| | This will open the Isaac Sim window and you can see the agent training in the environment. However, this |
| | will slow down the training process since the simulation is rendered on the screen. As a workaround, you |
| | can switch between different render modes in the ``"Isaac Lab"`` window that is docked on the bottom-right |
| | corner of the screen. To learn more about these render modes, please check the |
| | :class:`sim.SimulationContext.RenderMode` class. |
| |
|
| | Viewing the logs |
| | ~~~~~~~~~~~~~~~~ |
| |
|
| | On a separate terminal, you can monitor the training progress by executing the following command: |
| |
|
| | .. code:: bash |
| |
|
| | |
| | ./isaaclab.sh -p -m tensorboard.main |
| |
|
| | Playing the trained agent |
| | ~~~~~~~~~~~~~~~~~~~~~~~~~ |
| |
|
| | Once the training is complete, you can visualize the trained agent by executing the following command: |
| |
|
| | .. code:: bash |
| |
|
| | |
| | ./isaaclab.sh -p scripts/reinforcement_learning/sb3/play.py |
| |
|
| | The above command will load the latest checkpoint from the ``logs/sb3/Isaac-Cartpole-v0`` |
| | directory. You can also specify a specific checkpoint by passing the `` |
| |
|
| | .. _Stable-Baselines3: https://stable-baselines3.readthedocs.io/en/master/ |
| | .. _VecEnv API: https://stable-baselines3.readthedocs.io/en/master/guide/vec_envs.html |
| | .. _`stable_baselines3.common.vec_env.VecNormalize`: https://stable-baselines3.readthedocs.io/en/master/guide/vec_envs.html |
| | .. _RL-Games: https://github.com/Denys88/rl_games |
| | .. _RSL-RL: https://github.com/leggedrobotics/rsl_rl |
| | .. _SKRL: https://skrl.readthedocs.io |
| |
|