# {py:mod}`evorl.algorithms.ppo`

```{py:module} evorl.algorithms.ppo
```

```{autodoc2-docstring} evorl.algorithms.ppo
:parser: autodoc2_docstrings_parser
:allowtitles:
```

## Module Contents

### Classes

````{list-table}
:class: autosummary longtable
:align: left

* - {py:obj}`PPOAgent <evorl.algorithms.ppo.PPOAgent>`
  -
* - {py:obj}`PPONetworkParams <evorl.algorithms.ppo.PPONetworkParams>`
  - ```{autodoc2-docstring} evorl.algorithms.ppo.PPONetworkParams
    :parser: autodoc2_docstrings_parser
    :summary:
    ```
* - {py:obj}`PPOWorkflow <evorl.algorithms.ppo.PPOWorkflow>`
  -
````

### Functions

````{list-table}
:class: autosummary longtable
:align: left

* - {py:obj}`make_mlp_ppo_agent <evorl.algorithms.ppo.make_mlp_ppo_agent>`
  - ```{autodoc2-docstring} evorl.algorithms.ppo.make_mlp_ppo_agent
    :parser: autodoc2_docstrings_parser
    :summary:
    ```
````

### API

`````{py:class} PPOAgent
:canonical: evorl.algorithms.ppo.PPOAgent

Bases: {py:obj}`evorl.agent.Agent`

````{py:attribute} clip_epsilon
:canonical: evorl.algorithms.ppo.PPOAgent.clip_epsilon
:type: float
:value: >
   0.2

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.clip_epsilon
:parser: autodoc2_docstrings_parser
```

````

````{py:method} compute_actions(agent_state: evorl.agent.AgentState, sample_batch: evorl.sample_batch.SampleBatch, key: chex.PRNGKey) -> tuple[evorl.types.Action, evorl.types.PolicyExtraInfo]
:canonical: evorl.algorithms.ppo.PPOAgent.compute_actions

````

````{py:method} compute_values(agent_state: evorl.agent.AgentState, sample_batch: evorl.sample_batch.SampleBatch) -> chex.Array
:canonical: evorl.algorithms.ppo.PPOAgent.compute_values

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.compute_values
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} continuous_action
:canonical: evorl.algorithms.ppo.PPOAgent.continuous_action
:type: bool
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.continuous_action
:parser: autodoc2_docstrings_parser
```

````

````{py:method} evaluate_actions(agent_state: evorl.agent.AgentState, sample_batch: evorl.sample_batch.SampleBatch, key: chex.PRNGKey) -> tuple[evorl.types.Action, evorl.types.PolicyExtraInfo]
:canonical: evorl.algorithms.ppo.PPOAgent.evaluate_actions

````

````{py:method} init(obs_space: evorl.envs.Space, action_space: evorl.envs.Space, key: chex.PRNGKey) -> evorl.agent.AgentState
:canonical: evorl.algorithms.ppo.PPOAgent.init

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.init
:parser: autodoc2_docstrings_parser
```

````

````{py:method} loss(agent_state: evorl.agent.AgentState, sample_batch: evorl.sample_batch.SampleBatch, key: chex.PRNGKey) -> evorl.types.LossDict
:canonical: evorl.algorithms.ppo.PPOAgent.loss

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.loss
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} normalize_gae
:canonical: evorl.algorithms.ppo.PPOAgent.normalize_gae
:type: bool
:value: >
   True

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.normalize_gae
:parser: autodoc2_docstrings_parser
```

````

````{py:property} normalize_obs
:canonical: evorl.algorithms.ppo.PPOAgent.normalize_obs

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.normalize_obs
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} obs_preprocessor
:canonical: evorl.algorithms.ppo.PPOAgent.obs_preprocessor
:type: typing.Any
:value: >
   'pytree_field(...)'

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.obs_preprocessor
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} policy_network
:canonical: evorl.algorithms.ppo.PPOAgent.policy_network
:type: flax.linen.Module
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.policy_network
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} policy_obs_key
:canonical: evorl.algorithms.ppo.PPOAgent.policy_obs_key
:type: str
:value: <Multiline-String>

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.policy_obs_key
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} value_network
:canonical: evorl.algorithms.ppo.PPOAgent.value_network
:type: flax.linen.Module
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.value_network
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} value_obs_key
:canonical: evorl.algorithms.ppo.PPOAgent.value_obs_key
:type: str
:value: <Multiline-String>

```{autodoc2-docstring} evorl.algorithms.ppo.PPOAgent.value_obs_key
:parser: autodoc2_docstrings_parser
```

````

`````

`````{py:class} PPONetworkParams
:canonical: evorl.algorithms.ppo.PPONetworkParams

Bases: {py:obj}`evorl.types.PyTreeData`

```{autodoc2-docstring} evorl.algorithms.ppo.PPONetworkParams
:parser: autodoc2_docstrings_parser
```

````{py:attribute} policy_params
:canonical: evorl.algorithms.ppo.PPONetworkParams.policy_params
:type: evorl.types.Params
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.ppo.PPONetworkParams.policy_params
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} value_params
:canonical: evorl.algorithms.ppo.PPONetworkParams.value_params
:type: evorl.types.Params
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.ppo.PPONetworkParams.value_params
:parser: autodoc2_docstrings_parser
```

````

`````

`````{py:class} PPOWorkflow(env: evorl.envs.Env, agent: evorl.agent.Agent, optimizer: optax.GradientTransformation, evaluator: evorl.evaluators.Evaluator, config: omegaconf.DictConfig)
:canonical: evorl.algorithms.ppo.PPOWorkflow

Bases: {py:obj}`evorl.workflows.OnPolicyWorkflow`

````{py:method} learn(state: evorl.types.State) -> evorl.types.State
:canonical: evorl.algorithms.ppo.PPOWorkflow.learn

````

````{py:method} name()
:canonical: evorl.algorithms.ppo.PPOWorkflow.name
:classmethod:

````

````{py:method} step(state: evorl.types.State) -> tuple[evorl.metrics.MetricBase, evorl.types.State]
:canonical: evorl.algorithms.ppo.PPOWorkflow.step

````

`````

````{py:function} make_mlp_ppo_agent(action_space: evorl.envs.Space, clip_epsilon: float = 0.2, actor_hidden_layer_sizes: tuple[int] = (256, 256), critic_hidden_layer_sizes: tuple[int] = (256, 256), normalize_obs: bool = False, normalize_gae: bool = False, policy_obs_key: str = '', value_obs_key: str = '')
:canonical: evorl.algorithms.ppo.make_mlp_ppo_agent

```{autodoc2-docstring} evorl.algorithms.ppo.make_mlp_ppo_agent
:parser: autodoc2_docstrings_parser
```
````