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

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

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

## Module Contents

### Classes

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

* - {py:obj}`DQNAgent <evorl.algorithms.dqn.DQNAgent>`
  -
* - {py:obj}`DQNNetworkParams <evorl.algorithms.dqn.DQNNetworkParams>`
  -
* - {py:obj}`DQNTrainMetric <evorl.algorithms.dqn.DQNTrainMetric>`
  -
* - {py:obj}`DQNWorkflow <evorl.algorithms.dqn.DQNWorkflow>`
  -
* - {py:obj}`DQNWorkflowMetric <evorl.algorithms.dqn.DQNWorkflowMetric>`
  -
````

### Functions

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

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

### API

`````{py:class} DQNAgent
:canonical: evorl.algorithms.dqn.DQNAgent

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

````{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.dqn.DQNAgent.compute_actions

````

````{py:attribute} discount
:canonical: evorl.algorithms.dqn.DQNAgent.discount
:type: float
:value: >
   0.99

```{autodoc2-docstring} evorl.algorithms.dqn.DQNAgent.discount
: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.dqn.DQNAgent.evaluate_actions

````

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

```{autodoc2-docstring} evorl.algorithms.dqn.DQNAgent.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.dqn.DQNAgent.loss

```{autodoc2-docstring} evorl.algorithms.dqn.DQNAgent.loss
:parser: autodoc2_docstrings_parser
```

````

````{py:property} normalize_obs
:canonical: evorl.algorithms.dqn.DQNAgent.normalize_obs

```{autodoc2-docstring} evorl.algorithms.dqn.DQNAgent.normalize_obs
:parser: autodoc2_docstrings_parser
```

````

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

```{autodoc2-docstring} evorl.algorithms.dqn.DQNAgent.obs_preprocessor
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} q_network
:canonical: evorl.algorithms.dqn.DQNAgent.q_network
:type: flax.linen.Module
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.dqn.DQNAgent.q_network
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} target_type
:canonical: evorl.algorithms.dqn.DQNAgent.target_type
:type: str
:value: >
   'DDQN'

```{autodoc2-docstring} evorl.algorithms.dqn.DQNAgent.target_type
:parser: autodoc2_docstrings_parser
```

````

`````

`````{py:class} DQNNetworkParams
:canonical: evorl.algorithms.dqn.DQNNetworkParams

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

````{py:attribute} exploration_epsilon
:canonical: evorl.algorithms.dqn.DQNNetworkParams.exploration_epsilon
:type: float
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.dqn.DQNNetworkParams.exploration_epsilon
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} q_params
:canonical: evorl.algorithms.dqn.DQNNetworkParams.q_params
:type: evorl.types.Params
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.dqn.DQNNetworkParams.q_params
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} target_q_params
:canonical: evorl.algorithms.dqn.DQNNetworkParams.target_q_params
:type: evorl.types.Params
:value: >
   None

```{autodoc2-docstring} evorl.algorithms.dqn.DQNNetworkParams.target_q_params
:parser: autodoc2_docstrings_parser
```

````

`````

`````{py:class} DQNTrainMetric
:canonical: evorl.algorithms.dqn.DQNTrainMetric

Bases: {py:obj}`evorl.metrics.MetricBase`

````{py:attribute} loss
:canonical: evorl.algorithms.dqn.DQNTrainMetric.loss
:type: chex.Array
:value: >
   'zeros(...)'

```{autodoc2-docstring} evorl.algorithms.dqn.DQNTrainMetric.loss
:parser: autodoc2_docstrings_parser
```

````

````{py:attribute} raw_loss_dict
:canonical: evorl.algorithms.dqn.DQNTrainMetric.raw_loss_dict
:type: evorl.types.LossDict
:value: >
   'metric_field(...)'

```{autodoc2-docstring} evorl.algorithms.dqn.DQNTrainMetric.raw_loss_dict
:parser: autodoc2_docstrings_parser
```

````

`````

`````{py:class} DQNWorkflow(env: evorl.envs.Env, agent: evorl.agent.Agent, optimizer: optax.GradientTransformation, evaluator: evorl.evaluators.Evaluator, replay_buffer: evorl.replay_buffers.AbstractReplayBuffer, config: omegaconf.DictConfig)
:canonical: evorl.algorithms.dqn.DQNWorkflow

Bases: {py:obj}`evorl.algorithms.offpolicy_utils.OffPolicyWorkflowTemplate`

````{py:method} name()
:canonical: evorl.algorithms.dqn.DQNWorkflow.name
:classmethod:

````

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

````

`````

`````{py:class} DQNWorkflowMetric
:canonical: evorl.algorithms.dqn.DQNWorkflowMetric

Bases: {py:obj}`evorl.metrics.WorkflowMetric`

````{py:attribute} training_updates
:canonical: evorl.algorithms.dqn.DQNWorkflowMetric.training_updates
:type: chex.Array
:value: >
   'zeros(...)'

```{autodoc2-docstring} evorl.algorithms.dqn.DQNWorkflowMetric.training_updates
:parser: autodoc2_docstrings_parser
```

````

`````

````{py:function} make_mlp_discrete_dqn_agent(action_space: evorl.envs.Space, discount: float = 0.99, target_type: str = 'DDQN', q_hidden_layer_sizes: tuple[int] = (256, 256), normalize_obs: bool = False, value_obs_key: str = '')
:canonical: evorl.algorithms.dqn.make_mlp_discrete_dqn_agent

```{autodoc2-docstring} evorl.algorithms.dqn.make_mlp_discrete_dqn_agent
:parser: autodoc2_docstrings_parser
```
````