from typing import List, Tuple, Callable, Optional import torch import torch as th import torch.nn as nn from torchy_baselines.common.policies import BasePolicy, register_policy, create_mlp, BaseNetwork, \ create_sde_feature_extractor from torchy_baselines.common.distributions import StateDependentNoiseDistribution class Actor(BaseNetwork): """ Actor network (policy) for TD3. :param obs_dim: (int) Dimension of the observation :param action_dim: (int) Dimension of the action space :param net_arch: ([int]) Network architecture :param activation_fn: (nn.Module) Activation function :param use_sde: (bool) Whether to use State Dependent Exploration or not :param log_std_init: (float) Initial value for the log standard deviation :param clip_noise: (float) Clip the magnitude of the noise :param lr_sde: (float) Learning rate for the standard deviation of the noise :param full_std: (bool) Whether to use (n_features x n_actions) parameters for the std instead of only (n_features,) when using SDE. :param sde_net_arch: ([int]) Network architecture for extracting features when using SDE. If None, the latent features from the policy will be used. Pass an empty list to use the states as features. :param use_expln: (bool) Use `expln()` function instead of `exp()` when using SDE to ensure a positive standard deviation (cf paper). It allows to keep variance above zero and prevent it from growing too fast. In practice, `exp()` is usually enough. """ def __init__(self, obs_dim: int, action_dim: int, net_arch: List[int], activation_fn: nn.Module = nn.ReLU, use_sde: bool = False, log_std_init: float = -3, clip_noise: Optional[float] = None, lr_sde: float = 3e-4, full_std: bool = False, sde_net_arch: Optional[List[int]] = None, use_expln: bool = False): super(Actor, self).__init__() self.latent_pi, self.log_std = None, None self.weights_dist, self.exploration_mat = None, None self.use_sde, self.sde_optimizer = use_sde, None self.action_dim = action_dim self.full_std = full_std self.sde_feature_extractor = None if use_sde: latent_pi_net = create_mlp(obs_dim, -1, net_arch, activation_fn, squash_out=False) self.latent_pi = nn.Sequential(*latent_pi_net) latent_sde_dim = net_arch[-1] learn_features = sde_net_arch is not None # Separate feature extractor for SDE if sde_net_arch is not None: self.sde_feature_extractor, latent_sde_dim = create_sde_feature_extractor(obs_dim, sde_net_arch, activation_fn) # Create state dependent noise matrix (SDE) self.action_dist = StateDependentNoiseDistribution(action_dim, full_std=full_std, use_expln=use_expln, squash_output=False, learn_features=learn_features) action_net, self.log_std = self.action_dist.proba_distribution_net(latent_dim=net_arch[-1], latent_sde_dim=latent_sde_dim, log_std_init=log_std_init) # Squash output self.mu = nn.Sequential(action_net, nn.Tanh()) self.clip_noise = clip_noise self.sde_optimizer = th.optim.Adam([self.log_std], lr=lr_sde) self.reset_noise() else: actor_net = create_mlp(obs_dim, action_dim, net_arch, activation_fn, squash_out=True) self.mu = nn.Sequential(*actor_net) def get_std(self) -> torch.Tensor: """ Retrieve the standard deviation of the action distribution. Only useful when using SDE. It corresponds to `th.exp(log_std)` in the normal case, but is slightly different when using `expln` function (cf StateDependentNoiseDistribution doc). :return: (th.Tensor) """ return self.action_dist.get_std(self.log_std) def _get_action_dist_from_latent(self, latent_pi, latent_sde): mean_actions = self.mu(latent_pi) return self.action_dist.proba_distribution(mean_actions, self.log_std, latent_sde) def _get_latent(self, obs) -> Tuple[torch.Tensor, torch.Tensor]: latent_pi = self.latent_pi(obs) if self.sde_feature_extractor is not None: latent_sde = self.sde_feature_extractor(obs) else: latent_sde = latent_pi return latent_pi, latent_sde def evaluate_actions(self, obs: torch.Tensor, action: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]: """ Evaluate actions according to the current policy, given the observations. Only useful when using SDE. :param obs: (th.Tensor) :param action: (th.Tensor) :return: (th.Tensor, th.Tensor) log likelihood of taking those actions and entropy of the action distribution. """ latent_pi, latent_sde = self._get_latent(obs) _, distribution = self._get_action_dist_from_latent(latent_pi, latent_sde) log_prob = distribution.log_prob(action) # value = self.value_net(latent_vf) return log_prob, distribution.entropy() def reset_noise(self) -> None: """ Sample new weights for the exploration matrix, when using SDE. """ self.action_dist.sample_weights(self.log_std) def forward(self, obs: torch.Tensor, deterministic: bool = True) -> torch.Tensor: if self.use_sde: latent_pi, latent_sde = self._get_latent(obs) if deterministic: return self.mu(latent_pi) noise = self.action_dist.get_noise(latent_sde) if self.clip_noise is not None: noise = th.clamp(noise, -self.clip_noise, self.clip_noise) # TODO: Replace with squashing -> need to account for that in the sde update # -> set squash_out=True in the action_dist? # NOTE: the clipping is done in the rollout for now return self.mu(latent_pi) + noise # action, _ = self._get_action_dist_from_latent(latent_pi) # return action else: return self.mu(obs) class Critic(BaseNetwork): """ Critic network for TD3, in fact it represents the action-state value function (Q-value function) :param obs_dim: (int) Dimension of the observation :param action_dim: (int) Dimension of the action space :param net_arch: ([int]) Network architecture :param activation_fn: (nn.Module) Activation function """ def __init__(self, obs_dim: int, action_dim: int, net_arch: List[int], activation_fn: nn.Module = nn.ReLU): super(Critic, self).__init__() q1_net = create_mlp(obs_dim + action_dim, 1, net_arch, activation_fn) self.q1_net = nn.Sequential(*q1_net) q2_net = create_mlp(obs_dim + action_dim, 1, net_arch, activation_fn) self.q2_net = nn.Sequential(*q2_net) def forward(self, obs: torch.Tensor, action: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]: qvalue_input = th.cat([obs, action], dim=1) return self.q1_net(qvalue_input), self.q2_net(qvalue_input) def q1_forward(self, obs: torch.Tensor, action: torch.Tensor) -> torch.Tensor: return self.q1_net(th.cat([obs, action], dim=1)) class ValueFunction(BaseNetwork): """ Value function for TD3 when doing on-policy exploration with SDE. :param obs_dim: (int) Dimension of the observation :param net_arch: ([int]) Network architecture :param activation_fn: (nn.Module) Activation function """ def __init__(self, obs_dim, net_arch=None, activation_fn=nn.Tanh): super(ValueFunction, self).__init__() if net_arch is None: net_arch = [64, 64] vf_net = create_mlp(obs_dim, 1, net_arch, activation_fn) self.vf_net = nn.Sequential(*vf_net) def forward(self, obs): return self.vf_net(obs) class TD3Policy(BasePolicy): """ Policy class (with both actor and critic) for TD3. :param observation_space: (gym.spaces.Space) Observation space :param action_space: (gym.spaces.Space) Action space :param learning_rate: (callable) Learning rate schedule (could be constant) :param net_arch: ([int or dict]) The specification of the policy and value networks. :param device: (str or th.device) Device on which the code should run. :param activation_fn: (nn.Module) Activation function :param use_sde: (bool) Whether to use State Dependent Exploration or not :param log_std_init: (float) Initial value for the log standard deviation :param sde_net_arch: ([int]) Network architecture for extracting features when using SDE. If None, the latent features from the policy will be used. Pass an empty list to use the states as features. :param use_expln: (bool) Use `expln()` function instead of `exp()` when using SDE to ensure a positive standard deviation (cf paper). It allows to keep variance above zero and prevent it from growing too fast. In practice, `exp()` is usually enough. """ def __init__(self, observation_space, action_space, learning_rate, net_arch=None, device='cpu', activation_fn=nn.ReLU, use_sde=False, log_std_init=-3, clip_noise=None, lr_sde=3e-4, sde_net_arch=None, use_expln=False): super(TD3Policy, self).__init__(observation_space, action_space, device) # Default network architecture, from the original paper if net_arch is None: net_arch = [400, 300] self.obs_dim = self.observation_space.shape[0] self.action_dim = self.action_space.shape[0] self.net_arch = net_arch self.activation_fn = activation_fn self.net_args = { 'obs_dim': self.obs_dim, 'action_dim': self.action_dim, 'net_arch': self.net_arch, 'activation_fn': self.activation_fn } self.actor_kwargs = self.net_args.copy() sde_kwargs = { 'use_sde': use_sde, 'log_std_init': log_std_init, 'clip_noise': clip_noise, 'lr_sde': lr_sde, 'sde_net_arch': sde_net_arch, 'use_expln': use_expln } self.actor_kwargs.update(sde_kwargs) self.actor, self.actor_target = None, None self.critic, self.critic_target = None, None # For SDE only self.use_sde = use_sde self.vf_net = None self.log_std_init = log_std_init self._build(learning_rate) def _build(self, learning_rate): self.actor = self.make_actor() self.actor_target = self.make_actor() self.actor_target.load_state_dict(self.actor.state_dict()) self.actor.optimizer = th.optim.Adam(self.actor.parameters(), lr=learning_rate(1)) self.critic = self.make_critic() self.critic_target = self.make_critic() self.critic_target.load_state_dict(self.critic.state_dict()) self.critic.optimizer = th.optim.Adam(self.critic.parameters(), lr=learning_rate(1)) if self.use_sde: self.vf_net = ValueFunction(self.obs_dim) self.actor.sde_optimizer.add_param_group({'params': self.vf_net.parameters()}) def reset_noise(self): return self.actor.reset_noise() def make_actor(self): return Actor(**self.actor_kwargs).to(self.device) def make_critic(self): return Critic(**self.net_args).to(self.device) def forward(self, obs, deterministic=True): return self.actor(obs, deterministic=deterministic) MlpPolicy = TD3Policy register_policy("MlpPolicy", MlpPolicy)