Deep Deterministic Policy Gradient (DDPG)¶

DDPG is a model-free, deterministic off-policy actor-critic algorithm that uses deep function approximators to learn a policy (and to estimate the action-value function) in high-dimensional, continuous action spaces.

Paper: Continuous control with deep reinforcement learning.

Algorithm¶

Algorithm implementation¶

Main notation/symbols:

- policy function approximator (\(\mu_\theta\)), critic function approximator (\(Q_\phi\))
- states (\(s\)), actions (\(a\)), rewards (\(r\)), next states (\(s'\)), terminated (\(d_{_{end}}\)), truncated (\(d_{_{timeout}}\))
- loss (\(L\))

Decision making¶

act(...)
\(a \leftarrow \mu_\theta(s)\)
\(noise \leftarrow\) sample noise
\(scale \leftarrow (1 - \text{timestep} \;/\) timesteps \() \; (\) initial_scale \(-\) final_scale \() \;+\) final_scale
\(a \leftarrow \text{clip}(a + noise * scale, {a}_{Low}, {a}_{High})\)

Learning algorithm¶

_update(...)

# gradient steps

FOR each gradient step up to gradient_steps DO

# sample a batch from memory
[\(s, a, r, s', d_{_{end}}, d_{_{timeout}}\)] with size batch_size
# compute target values
\(a' \leftarrow \mu_{\theta_{target}}(s')\)
\(Q_{_{target}} \leftarrow Q_{\phi_{target}}(s', a')\)
\(y \leftarrow r \;+\) discount_factor \(\neg (d_{_{end}} \lor d_{_{timeout}}) \; Q_{_{target}}\)
# compute critic loss
\(Q \leftarrow Q_\phi(s, a)\)
\(L_{Q_\phi} \leftarrow \frac{1}{N} \sum_{i=1}^N (Q - y)^2\)
# optimization step (critic)
reset \(\text{optimizer}_\phi\)
\(\nabla_{\phi} L_{Q_\phi}\)
\(\text{clip}(\lVert \nabla_{\phi} \rVert)\) with grad_norm_clip
step \(\text{optimizer}_\phi\)
# compute policy (actor) loss
\(a \leftarrow \mu_\theta(s)\)
\(Q \leftarrow Q_\phi(s, a)\)
\(L_{\mu_\theta} \leftarrow - \frac{1}{N} \sum_{i=1}^N Q\)
# optimization step (policy)
reset \(\text{optimizer}_\theta\)
\(\nabla_{\theta} L_{\mu_\theta}\)
\(\text{clip}(\lVert \nabla_{\theta} \rVert)\) with grad_norm_clip
step \(\text{optimizer}_\theta\)
# update target networks
\(\theta_{target} \leftarrow\) polyak \(\theta + (1 \;-\) polyak \() \theta_{target}\)
\(\phi_{target} \leftarrow\) polyak \(\phi + (1 \;-\) polyak \() \phi_{target}\)
# update learning rate
IF there is a learning_rate_scheduler THEN
step \(\text{scheduler}_\theta (\text{optimizer}_\theta)\)
step \(\text{scheduler}_\phi (\text{optimizer}_\phi)\)

Usage¶

Note

Support for recurrent neural networks (RNN, LSTM, GRU and any other variant) is implemented in a separate file (ddpg_rnn.py) to maintain the readability of the standard implementation (ddpg.py).

# import the agent and its default configuration
from skrl.agents.torch.ddpg import DDPG, DDPG_CFG

# instantiate the agent's models
models = {}
models["policy"] = ...
models["target_policy"] = ...  # only required during training
models["critic"] = ...  # only required during training
models["target_critic"] = ...  # only required during training

# adjust some configuration if necessary
cfg_agent = DDPG_CFG()
cfg_agent.KEY = ...

# instantiate the agent
# (assuming a defined environment <env> and memory <memory>)
agent = DDPG(
    models=models,
    memory=memory,  # only required during training
    cfg=cfg_agent,
    observation_space=env.observation_space,
    state_space=env.state_space,
    action_space=env.action_space,
    device=env.device,
)

# import the agent and its default configuration
from skrl.agents.jax.ddpg import DDPG, DDPG_CFG

# instantiate the agent's models
models = {}
models["policy"] = ...
models["target_policy"] = ...  # only required during training
models["critic"] = ...  # only required during training
models["target_critic"] = ...  # only required during training

# adjust some configuration if necessary
cfg_agent = DDPG_CFG()
cfg_agent.KEY = ...

# instantiate the agent
# (assuming a defined environment <env> and memory <memory>)
agent = DDPG(
    models=models,
    memory=memory,  # only required during training
    cfg=cfg_agent,
    observation_space=env.observation_space,
    state_space=env.state_space,
    action_space=env.action_space,
    device=env.device,
)

Note

When using recursive models it is necessary to override their .get_specification() method. Visit each model’s documentation for more details.

# import the agent and its default configuration
from skrl.agents.torch.ddpg import DDPG_RNN as DDPG, DDPG_CFG

# instantiate the agent's models
models = {}
models["policy"] = ...
models["target_policy"] = ...  # only required during training
models["critic"] = ...  # only required during training
models["target_critic"] = ...  # only required during training

# adjust some configuration if necessary
cfg_agent = DDPG_CFG()
cfg_agent.KEY = ...

# instantiate the agent
# (assuming a defined environment <env> and memory <memory>)
agent = DDPG(
    models=models,
    memory=memory,  # only required during training
    cfg=cfg_agent,
    observation_space=env.observation_space,
    state_space=env.state_space,
    action_space=env.action_space,
    device=env.device,
)

Configuration and hyperparameters¶

Dataclass
`DDPG_CFG`	`DDPG_CFG`	`DDPG_CFG`	`DDPG_CFG`

Spaces¶

The implementation supports the following Gymnasium spaces:

Gymnasium spaces	Observation	Action
Discrete	\(\square\)	\(\square\)
MultiDiscrete	\(\square\)	\(\square\)
Box	\(\blacksquare\)	\(\blacksquare\)
Dict	\(\blacksquare\)	\(\square\)

Models¶

The implementation uses 4 deterministic function approximators. These function approximators (models) must be collected in a dictionary and passed to the constructor of the class under the argument models.

Notation	Concept	Key	Input shape	Output shape	Type
\(\mu_\theta(s)\)	Policy (actor)	`"policy"`	observation	action	Deterministic
\(\mu_{\theta_{target}}(s)\)	Target policy	`"target_policy"`	observation	action	Deterministic
\(Q_\phi(s, a)\)	Q-network (critic)	`"critic"`	observation + action	1	Deterministic
\(Q_{\phi_{target}}(s, a)\)	Target Q-network	`"target_critic"`	observation + action	1	Deterministic

Features¶

Support for advanced features is described in the following table:

Feature	Support and remarks
Shared model	-	\(\square\)	\(\square\)	\(\square\)
RNN support	RNN, LSTM, GRU and any other variant	\(\blacksquare\)	\(\square\)	\(\square\)
Mixed precision	Automatic mixed precision	\(\blacksquare\)	\(\square\)	\(\square\)
Distributed	Single Program Multi Data (SPMD) multi-GPU	\(\blacksquare\)	\(\blacksquare\)	\(\square\)

API¶

PyTorch¶

`DDPG_CFG`	Configuration for the DDPG agent.
`DDPG`	Deep Deterministic Policy Gradient (DDPG).
`DDPG_RNN`	Deep Deterministic Policy Gradient (DDPG) with support for Recurrent Neural Networks (RNN, GRU, LSTM, etc.).

class skrl.agents.torch.ddpg.DDPG_CFG(*, experiment: ExperimentCfg = <factory>, gradient_steps: int = 1, batch_size: int = 64, discount_factor: float = 0.99, polyak: float = 0.005, learning_rate: float | tuple[float, float] = 0.001, learning_rate_scheduler: type | tuple[type | None, type | None] | None = None, learning_rate_scheduler_kwargs: dict | tuple[dict, dict] = <factory>, observation_preprocessor: type | None = None, observation_preprocessor_kwargs: dict = <factory>, state_preprocessor: type | None = None, state_preprocessor_kwargs: dict = <factory>, random_timesteps: int = 0, learning_starts: int = 0, grad_norm_clip: float = 0, exploration_noise: type | None = None, exploration_noise_kwargs: dict = <factory>, exploration_scheduler: Callable[[int, int], float] | None = None, rewards_shaper: Callable | None = None, mixed_precision: bool = False)[source]¶

Bases: AgentCfg

Configuration for the DDPG agent.

Methods:

`expand`()	Expand the configuration.
`validate`()	Validate the configuration.

Attributes:

`batch_size`	Batch size for sampling transitions from memory during training.
`discount_factor`	Parameter that balances the importance of future rewards (close to 1.0) versus immediate rewards (close to 0.0).
`experiment`	Experiment settings.
`exploration_noise`	Exploration noise class.
`exploration_noise_kwargs`	Keyword arguments for the exploration noise's constructor.
`exploration_scheduler`	Exploration scheduler function.
`grad_norm_clip`	Clipping coefficient for the gradients by their global norm.
`gradient_steps`	Number of gradient steps to perform for each update.
`learning_rate`	Learning rate for the actor and critic networks.
`learning_rate_scheduler`	Learning rate scheduler class for the actor and critic networks.
`learning_rate_scheduler_kwargs`	Keyword arguments for the learning rate scheduler's constructor.
`learning_starts`	Number of steps to perform before calling the algorithm update function.
`mixed_precision`	Whether to enable automatic mixed precision for higher performance.
`observation_preprocessor`	Preprocessor class to process the environment's observations.
`observation_preprocessor_kwargs`	Keyword arguments for the observation preprocessor's constructor.
`polyak`	Parameter to control the update of the target networks by polyak averaging.
`random_timesteps`	Number of random exploration (sampling random actions) steps to perform before sampling actions from the policy.
`rewards_shaper`	Rewards shaping function.
`state_preprocessor`	Preprocessor class to process the environment's states.
`state_preprocessor_kwargs`	Keyword arguments for the state preprocessor's constructor.

expand() → None[source]¶: Expand the configuration.

validate() → bool[source]¶: Validate the configuration.

batch_size: int = 64¶: Batch size for sampling transitions from memory during training.

discount_factor: float = 0.99¶

Parameter that balances the importance of future rewards (close to 1.0) versus immediate rewards (close to 0.0).

Range: [0.0, 1.0].

experiment: ExperimentCfg¶: Experiment settings.

exploration_noise: type | None = None¶

Exploration noise class.

See Noises for more details.

exploration_noise_kwargs: dict¶

Keyword arguments for the exploration noise’s constructor.

See Noises for more details.

exploration_scheduler: Callable[[int, int], float] | None = None¶

Exploration scheduler function.

The function takes the current timestep and the total number of timesteps as arguments and returns a scale value to apply to the sampled exploration noise.

grad_norm_clip: float = 0¶

Clipping coefficient for the gradients by their global norm.

If less than or equal to 0, the gradients will not be clipped.

gradient_steps: int = 1¶: Number of gradient steps to perform for each update.

learning_rate: float | tuple[float, float] = 0.001¶

Learning rate for the actor and critic networks.

If a float is provided, the same learning rate will be used for the networks.
If a tuple is provided, its elements will be used for each network in order.

learning_rate_scheduler: type | tuple[type | None, type | None] | None = None¶

Learning rate scheduler class for the actor and critic networks.

See Learning rate schedulers for more details.

If a class is provided, the same learning rate scheduler will be used for the networks.
If a tuple is provided, its elements will be used for each network in order.

learning_rate_scheduler_kwargs: dict | tuple[dict, dict]¶

Keyword arguments for the learning rate scheduler’s constructor.

See Learning rate schedulers for more details.

Warning

The optimizer argument is automatically passed to the learning rate scheduler’s constructor. Therefore, it must not be provided in the keyword arguments.

If a dictionary is provided, the same keyword arguments will be used for the networks.
If a tuple is provided, its elements will be used for each network in order.

learning_starts: int = 0¶: Number of steps to perform before calling the algorithm update function.

mixed_precision: bool = False¶: Whether to enable automatic mixed precision for higher performance.

observation_preprocessor: type | None = None¶

Preprocessor class to process the environment’s observations.

See Preprocessors for more details.

observation_preprocessor_kwargs: dict¶

Keyword arguments for the observation preprocessor’s constructor.

See Preprocessors for more details.

polyak: float = 0.005¶

Parameter to control the update of the target networks by polyak averaging.

Range: [0.0, 1.0]. See update_parameters() for more details.

random_timesteps: int = 0¶: Number of random exploration (sampling random actions) steps to perform before sampling actions from the policy.

rewards_shaper: Callable | None = None¶: Rewards shaping function.

state_preprocessor: type | None = None¶

Preprocessor class to process the environment’s states.

See Preprocessors for more details.

state_preprocessor_kwargs: dict¶

Keyword arguments for the state preprocessor’s constructor.

See Preprocessors for more details.

class skrl.agents.torch.ddpg.DDPG(*, models: dict[str, Model], memory: Memory | None = None, observation_space: gymnasium.Space | None = None, state_space: gymnasium.Space | None = None, action_space: gymnasium.Space | None = None, device: str | torch.device | None = None, cfg: DDPG_CFG | dict = {})[source]¶

Bases: Agent

Deep Deterministic Policy Gradient (DDPG).

https://arxiv.org/abs/1509.02971

Parameters:

models – Agent’s models.
memory – Memory to storage agent’s data and environment transitions.
observation_space – Observation space.
state_space – State space.
action_space – Action space.
device – Data allocation and computation device. If not specified, the default device will be used.
cfg – Agent’s configuration.

Raises:

KeyError – If a configuration key is missing.

Methods:

`act`(observations, states, *, timestep, timesteps)	Process the environment's observations/states to make a decision (actions) using the main policy.
`enable_models_training_mode`([enabled])	Set the training mode of all the agent's models: enabled (training) or disabled (evaluation).
`enable_training_mode`([enabled, apply_to_models])	Set the training mode of the agent: enabled (training) or disabled (evaluation).
`init`(*[, trainer_cfg])	Initialize the agent.
`load`(path)	Load the agent from the specified path.
`post_interaction`(*, timestep, timesteps)	Method called after the interaction with the environment.
`pre_interaction`(*, timestep, timesteps)	Method called before the interaction with the environment.
`record_transition`(*, observations, states, ...)	Record an environment transition in memory.
`save`(path)	Save the agent to the specified path.
`track_data`(tag, value)	Track data to TensorBoard.
`update`(*, timestep, timesteps)	Algorithm's main update step.
`write_checkpoint`(*, timestep, timesteps)	Write checkpoint (modules) to persistent storage.
`write_tracking_data`(*, timestep, timesteps)	Write tracking data to TensorBoard.

act(observations: torch.Tensor, states: torch.Tensor | None, *, timestep: int, timesteps: int) → tuple[torch.Tensor, dict[str, Any]][source]¶

Process the environment’s observations/states to make a decision (actions) using the main policy.

Parameters:

observations – Environment observations.
states – Environment states.
timestep – Current timestep.
timesteps – Number of timesteps.

Returns:

Agent output. The first component is the expected action/value returned by the agent. The second component is a dictionary containing extra output values according to the model.

enable_models_training_mode(enabled: bool = True) → None[source]¶

Set the training mode of all the agent’s models: enabled (training) or disabled (evaluation).

Parameters:: enabled – True to enable the training mode, False to enable the evaluation mode.

enable_training_mode(enabled: bool = True, *, apply_to_models: bool = False) → None[source]¶

Set the training mode of the agent: enabled (training) or disabled (evaluation).

The training mode can be queried by the training property.

Parameters:

enabled – True to enable the training mode, False to enable the evaluation mode.
apply_to_models – Whether to apply the training mode to all the agent’s models.

init(*, trainer_cfg: dict[str, Any] | None = None) → None[source]¶

Initialize the agent.

Parameters:: trainer_cfg – Trainer configuration.

load(path: str) → None[source]¶

Load the agent from the specified path.

Note

The final storage device is determined by the constructor of the agent.

Parameters:: path – Path to load the agent from.

post_interaction(*, timestep: int, timesteps: int) → None[source]¶

Method called after the interaction with the environment.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

pre_interaction(*, timestep: int, timesteps: int) → None[source]¶

Method called before the interaction with the environment.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

record_transition(*, observations: torch.Tensor, states: torch.Tensor, actions: torch.Tensor, rewards: torch.Tensor, next_observations: torch.Tensor, next_states: torch.Tensor, terminated: torch.Tensor, truncated: torch.Tensor, infos: Any, timestep: int, timesteps: int) → None[source]¶

Record an environment transition in memory.

Parameters:

observations – Environment observations.
states – Environment states.
actions – Actions taken by the agent.
rewards – Instant rewards achieved by the current actions.
next_observations – Next environment observations.
next_states – Next environment states.
terminated – Signals that indicate episodes have terminated.
truncated – Signals that indicate episodes have been truncated.
infos – Additional information about the environment.
timestep – Current timestep.
timesteps – Number of timesteps.

save(path: str) → None[source]¶

Save the agent to the specified path.

Parameters:: path – Path to save the agent to.

track_data(tag: str, value: float) → None[source]¶

Track data to TensorBoard.

Note

Currently only scalar data is supported.

Parameters:

tag – Data identifier (e.g. ‘Loss/Policy loss’).
value – Value to track.

update(*, timestep: int, timesteps: int) → None[source]¶

Algorithm’s main update step.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

write_checkpoint(*, timestep: int, timesteps: int) → None[source]¶

Write checkpoint (modules) to persistent storage.

Note

The checkpoints are stored in the subdirectory checkpoints within the experiment directory. The checkpoint name is the timestep argument value (if it is not None), or the current system date-time otherwise.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

write_tracking_data(*, timestep: int, timesteps: int) → None[source]¶

Write tracking data to TensorBoard.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

class skrl.agents.torch.ddpg.DDPG_RNN(*, models: dict[str, Model], memory: Memory | None = None, observation_space: gymnasium.Space | None = None, state_space: gymnasium.Space | None = None, action_space: gymnasium.Space | None = None, device: str | torch.device | None = None, cfg: DDPG_CFG | dict = {})[source]¶

Bases: Agent

Deep Deterministic Policy Gradient (DDPG) with support for Recurrent Neural Networks (RNN, GRU, LSTM, etc.).

https://arxiv.org/abs/1509.02971

Parameters:

models – Agent’s models.
memory – Memory to storage agent’s data and environment transitions.
observation_space – Observation space.
state_space – State space.
action_space – Action space.
device – Data allocation and computation device. If not specified, the default device will be used.
cfg – Agent’s configuration.

Raises:

KeyError – If a configuration key is missing.

Methods:

`act`(observations, states, *, timestep, timesteps)	Process the environment's observations/states to make a decision (actions) using the main policy.
`enable_models_training_mode`([enabled])	Set the training mode of all the agent's models: enabled (training) or disabled (evaluation).
`enable_training_mode`([enabled, apply_to_models])	Set the training mode of the agent: enabled (training) or disabled (evaluation).
`init`(*[, trainer_cfg])	Initialize the agent.
`load`(path)	Load the agent from the specified path.
`post_interaction`(*, timestep, timesteps)	Method called after the interaction with the environment.
`pre_interaction`(*, timestep, timesteps)	Method called before the interaction with the environment.
`record_transition`(*, observations, states, ...)	Record an environment transition in memory.
`save`(path)	Save the agent to the specified path.
`track_data`(tag, value)	Track data to TensorBoard.
`update`(*, timestep, timesteps)	Algorithm's main update step.
`write_checkpoint`(*, timestep, timesteps)	Write checkpoint (modules) to persistent storage.
`write_tracking_data`(*, timestep, timesteps)	Write tracking data to TensorBoard.

act(observations: torch.Tensor, states: torch.Tensor | None, *, timestep: int, timesteps: int) → tuple[torch.Tensor, dict[str, Any]][source]¶

Process the environment’s observations/states to make a decision (actions) using the main policy.

Parameters:

observations – Environment observations.
states – Environment states.
timestep – Current timestep.
timesteps – Number of timesteps.

Returns:

Agent output. The first component is the expected action/value returned by the agent. The second component is a dictionary containing extra output values according to the model.

enable_models_training_mode(enabled: bool = True) → None[source]¶

Set the training mode of all the agent’s models: enabled (training) or disabled (evaluation).

Parameters:: enabled – True to enable the training mode, False to enable the evaluation mode.

enable_training_mode(enabled: bool = True, *, apply_to_models: bool = False) → None[source]¶

Set the training mode of the agent: enabled (training) or disabled (evaluation).

The training mode can be queried by the training property.

Parameters:

enabled – True to enable the training mode, False to enable the evaluation mode.
apply_to_models – Whether to apply the training mode to all the agent’s models.

init(*, trainer_cfg: dict[str, Any] | None = None) → None[source]¶

Initialize the agent.

Parameters:: trainer_cfg – Trainer configuration.

load(path: str) → None[source]¶

Load the agent from the specified path.

Note

The final storage device is determined by the constructor of the agent.

Parameters:: path – Path to load the agent from.

post_interaction(*, timestep: int, timesteps: int) → None[source]¶

Method called after the interaction with the environment.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

pre_interaction(*, timestep: int, timesteps: int) → None[source]¶

Method called before the interaction with the environment.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

Record an environment transition in memory.

Parameters:

observations – Environment observations.
states – Environment states.
actions – Actions taken by the agent.
rewards – Instant rewards achieved by the current actions.
next_observations – Next environment observations.
next_states – Next environment states.
terminated – Signals that indicate episodes have terminated.
truncated – Signals that indicate episodes have been truncated.
infos – Additional information about the environment.
timestep – Current timestep.
timesteps – Number of timesteps.

save(path: str) → None[source]¶

Save the agent to the specified path.

Parameters:: path – Path to save the agent to.

track_data(tag: str, value: float) → None[source]¶

Track data to TensorBoard.

Note

Currently only scalar data is supported.

Parameters:

tag – Data identifier (e.g. ‘Loss/Policy loss’).
value – Value to track.

update(*, timestep: int, timesteps: int) → None[source]¶

Algorithm’s main update step.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

write_checkpoint(*, timestep: int, timesteps: int) → None[source]¶

Write checkpoint (modules) to persistent storage.

Note

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

write_tracking_data(*, timestep: int, timesteps: int) → None[source]¶

Write tracking data to TensorBoard.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

JAX¶

`DDPG_CFG`	Configuration for the DDPG agent.
`DDPG`	Deep Deterministic Policy Gradient (DDPG).

class skrl.agents.jax.ddpg.DDPG_CFG(*, experiment: ExperimentCfg = <factory>, gradient_steps: int = 1, batch_size: int = 64, discount_factor: float = 0.99, polyak: float = 0.005, learning_rate: float | tuple[float, float] = 0.001, learning_rate_scheduler: type | tuple[type | None, type | None] | None = None, learning_rate_scheduler_kwargs: dict | tuple[dict, dict] = <factory>, observation_preprocessor: type | None = None, observation_preprocessor_kwargs: dict = <factory>, state_preprocessor: type | None = None, state_preprocessor_kwargs: dict = <factory>, random_timesteps: int = 0, learning_starts: int = 0, grad_norm_clip: float = 0, exploration_noise: type | None = None, exploration_noise_kwargs: dict = <factory>, exploration_scheduler: Callable[[int, int], float] | None = None, rewards_shaper: Callable | None = None)[source]¶

Bases: AgentCfg

Configuration for the DDPG agent.

Methods:

`expand`()	Expand the configuration.
`validate`()	Validate the configuration.

Attributes:

`batch_size`	Batch size for sampling transitions from memory during training.
`discount_factor`	Parameter that balances the importance of future rewards (close to 1.0) versus immediate rewards (close to 0.0).
`experiment`	Experiment settings.
`exploration_noise`	Exploration noise class.
`exploration_noise_kwargs`	Keyword arguments for the exploration noise's constructor.
`exploration_scheduler`	Exploration scheduler function.
`grad_norm_clip`	Clipping coefficient for the gradients by their global norm.
`gradient_steps`	Number of gradient steps to perform for each update.
`learning_rate`	Learning rate for the actor and critic networks.
`learning_rate_scheduler`	Learning rate scheduler class for the actor and critic networks.
`learning_rate_scheduler_kwargs`	Keyword arguments for the learning rate scheduler's constructor.
`learning_starts`	Number of steps to perform before calling the algorithm update function.
`observation_preprocessor`	Preprocessor class to process the environment's observations.
`observation_preprocessor_kwargs`	Keyword arguments for the observation preprocessor's constructor.
`polyak`	Parameter to control the update of the target networks by polyak averaging.
`random_timesteps`	Number of random exploration (sampling random actions) steps to perform before sampling actions from the policy.
`rewards_shaper`	Rewards shaping function.
`state_preprocessor`	Preprocessor class to process the environment's states.
`state_preprocessor_kwargs`	Keyword arguments for the state preprocessor's constructor.

expand() → None[source]¶: Expand the configuration.

validate() → bool[source]¶: Validate the configuration.

batch_size: int = 64¶: Batch size for sampling transitions from memory during training.

discount_factor: float = 0.99¶

Parameter that balances the importance of future rewards (close to 1.0) versus immediate rewards (close to 0.0).

Range: [0.0, 1.0].

experiment: ExperimentCfg¶: Experiment settings.

exploration_noise: type | None = None¶

Exploration noise class.

See Noises for more details.

exploration_noise_kwargs: dict¶

Keyword arguments for the exploration noise’s constructor.

See Noises for more details.

exploration_scheduler: Callable[[int, int], float] | None = None¶

Exploration scheduler function.

The function takes the current timestep and the total number of timesteps as arguments and returns a scale value to apply to the sampled exploration noise.

grad_norm_clip: float = 0¶

Clipping coefficient for the gradients by their global norm.

If less than or equal to 0, the gradients will not be clipped.

gradient_steps: int = 1¶: Number of gradient steps to perform for each update.

learning_rate: float | tuple[float, float] = 0.001¶

Learning rate for the actor and critic networks.

If a float is provided, the same learning rate will be used for the networks.
If a tuple is provided, its elements will be used for each network in order.

learning_rate_scheduler: type | tuple[type | None, type | None] | None = None¶

Learning rate scheduler class for the actor and critic networks.

See Learning rate schedulers for more details.

If a class is provided, the same learning rate scheduler will be used for the networks.
If a tuple is provided, its elements will be used for each network in order.

learning_rate_scheduler_kwargs: dict | tuple[dict, dict]¶

Keyword arguments for the learning rate scheduler’s constructor.

See Learning rate schedulers for more details.

Warning

The optimizer argument is automatically passed to the learning rate scheduler’s constructor. Therefore, it must not be provided in the keyword arguments.

If a dictionary is provided, the same keyword arguments will be used for the networks.
If a tuple is provided, its elements will be used for each network in order.

learning_starts: int = 0¶: Number of steps to perform before calling the algorithm update function.

observation_preprocessor: type | None = None¶

Preprocessor class to process the environment’s observations.

See Preprocessors for more details.

observation_preprocessor_kwargs: dict¶

Keyword arguments for the observation preprocessor’s constructor.

See Preprocessors for more details.

polyak: float = 0.005¶

Parameter to control the update of the target networks by polyak averaging.

Range: [0.0, 1.0]. See update_parameters() for more details.

random_timesteps: int = 0¶: Number of random exploration (sampling random actions) steps to perform before sampling actions from the policy.

rewards_shaper: Callable | None = None¶: Rewards shaping function.

state_preprocessor: type | None = None¶

Preprocessor class to process the environment’s states.

See Preprocessors for more details.

state_preprocessor_kwargs: dict¶

Keyword arguments for the state preprocessor’s constructor.

See Preprocessors for more details.

class skrl.agents.jax.ddpg.DDPG(*, models: dict[str, Model], memory: Memory | None = None, observation_space: gymnasium.Space | None = None, state_space: gymnasium.Space | None = None, action_space: gymnasium.Space | None = None, device: str | jax.Device | None = None, cfg: DDPG_CFG | dict = {})[source]¶

Bases: Agent

Deep Deterministic Policy Gradient (DDPG).

https://arxiv.org/abs/1509.02971

Parameters:

models – Agent’s models.
memory – Memory to storage agent’s data and environment transitions.
observation_space – Observation space.
state_space – State space.
action_space – Action space.
device – Data allocation and computation device. If not specified, the default device will be used.
cfg – Agent’s configuration.

Raises:

KeyError – If a configuration key is missing.

Methods:

`act`(observations, states, *, timestep, timesteps)	Process the environment's observations/states to make a decision (actions) using the main policy.
`enable_models_training_mode`([enabled])	Set the training mode of all the agent's models: enabled (training) or disabled (evaluation).
`enable_training_mode`([enabled, apply_to_models])	Set the training mode of the agent: enabled (training) or disabled (evaluation).
`init`(*[, trainer_cfg])	Initialize the agent.
`load`(path)	Load the agent from the specified path.
`post_interaction`(*, timestep, timesteps)	Method called after the interaction with the environment.
`pre_interaction`(*, timestep, timesteps)	Method called before the interaction with the environment.
`record_transition`(*, observations, states, ...)	Record an environment transition in memory.
`save`(path)	Save the agent to the specified path.
`track_data`(tag, value)	Track data to TensorBoard.
`update`(*, timestep, timesteps)	Algorithm's main update step.
`write_checkpoint`(*, timestep, timesteps)	Write checkpoint (modules) to persistent storage.
`write_tracking_data`(*, timestep, timesteps)	Write tracking data to TensorBoard.

act(observations: jax.Array, states: jax.Array | None, *, timestep: int, timesteps: int) → tuple[jax.Array, dict[str, Any]][source]¶

Process the environment’s observations/states to make a decision (actions) using the main policy.

Parameters:

observations – Environment observations.
states – Environment states.
timestep – Current timestep.
timesteps – Number of timesteps.

Returns:

Agent output. The first component is the expected action/value returned by the agent. The second component is a dictionary containing extra output values according to the model.

enable_models_training_mode(enabled: bool = True) → None[source]¶

Set the training mode of all the agent’s models: enabled (training) or disabled (evaluation).

Parameters:: enabled – True to enable the training mode, False to enable the evaluation mode.

enable_training_mode(enabled: bool = True, *, apply_to_models: bool = False) → None[source]¶

Set the training mode of the agent: enabled (training) or disabled (evaluation).

The training mode can be queried by the training property.

Parameters:

enabled – True to enable the training mode, False to enable the evaluation mode.
apply_to_models – Whether to apply the training mode to all the agent’s models.

init(*, trainer_cfg: dict[str, Any] | None = None) → None[source]¶

Initialize the agent.

Parameters:: trainer_cfg – Trainer configuration.

load(path: str) → None[source]¶

Load the agent from the specified path.

Parameters:: path – Path to load the agent from.

post_interaction(*, timestep: int, timesteps: int) → None[source]¶

Method called after the interaction with the environment.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

pre_interaction(*, timestep: int, timesteps: int) → None[source]¶

Method called before the interaction with the environment.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

record_transition(*, observations: jax.Array, states: jax.Array, actions: jax.Array, rewards: jax.Array, next_observations: jax.Array, next_states: jax.Array, terminated: jax.Array, truncated: jax.Array, infos: Any, timestep: int, timesteps: int) → None[source]¶

Record an environment transition in memory.

Parameters:

observations – Environment observations.
states – Environment states.
actions – Actions taken by the agent.
rewards – Instant rewards achieved by the current actions.
next_observations – Next environment observations.
next_states – Next environment states.
terminated – Signals that indicate episodes have terminated.
truncated – Signals that indicate episodes have been truncated.
infos – Additional information about the environment.
timestep – Current timestep.
timesteps – Number of timesteps.

save(path: str) → None[source]¶

Save the agent to the specified path.

Parameters:: path – Path to save the agent to.

track_data(tag: str, value: float) → None[source]¶

Track data to TensorBoard.

Note

Currently only scalar data is supported.

Parameters:

tag – Data identifier (e.g. ‘Loss/Policy loss’).
value – Value to track.

update(*, timestep: int, timesteps: int) → None[source]¶

Algorithm’s main update step.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

write_checkpoint(*, timestep: int, timesteps: int) → None[source]¶

Write checkpoint (modules) to persistent storage.

Note

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

write_tracking_data(*, timestep: int, timesteps: int) → None[source]¶

Write tracking data to TensorBoard.

Parameters:

timestep – Current timestep.
timesteps – Number of timesteps.

Warp¶

`DDPG_CFG`	Configuration for the DDPG agent.
`DDPG`	Deep Deterministic Policy Gradient (DDPG).