from typing import Any, Dict, List, Optional, Type
import torch as th
from gymnasium import spaces
from stable_baselines3.common.policies import BasePolicy
from stable_baselines3.common.torch_layers import (
BaseFeaturesExtractor,
CombinedExtractor,
FlattenExtractor,
NatureCNN,
create_mlp,
)
from stable_baselines3.common.type_aliases import PyTorchObs, Schedule
from torch import nn
class QuantileNetwork(BasePolicy):
"""
Quantile network for QR-DQN
:param observation_space: Observation space
:param action_space: Action space
:param n_quantiles: Number of quantiles
:param net_arch: The specification of the network architecture.
:param activation_fn: Activation function
:param normalize_images: Whether to normalize images or not,
dividing by 255.0 (True by default)
"""
action_space: spaces.Discrete
def __init__(
self,
observation_space: spaces.Space,
action_space: spaces.Discrete,
features_extractor: BaseFeaturesExtractor,
features_dim: int,
n_quantiles: int = 200,
net_arch: Optional[List[int]] = None,
activation_fn: Type[nn.Module] = nn.ReLU,
normalize_images: bool = True,
):
super().__init__(
observation_space,
action_space,
features_extractor=features_extractor,
normalize_images=normalize_images,
)
if net_arch is None:
net_arch = [64, 64]
self.net_arch = net_arch
self.activation_fn = activation_fn
self.features_dim = features_dim
self.n_quantiles = n_quantiles
action_dim = int(self.action_space.n) # number of actions
quantile_net = create_mlp(self.features_dim, action_dim * self.n_quantiles, self.net_arch, self.activation_fn)
self.quantile_net = nn.Sequential(*quantile_net)
def forward(self, obs: PyTorchObs) -> th.Tensor:
"""
Predict the quantiles.
:param obs: Observation
:return: The estimated quantiles for each action.
"""
quantiles = self.quantile_net(self.extract_features(obs, self.features_extractor))
return quantiles.view(-1, self.n_quantiles, int(self.action_space.n))
def _predict(self, observation: PyTorchObs, deterministic: bool = True) -> th.Tensor:
q_values = self(observation).mean(dim=1)
# Greedy action
action = q_values.argmax(dim=1).reshape(-1)
return action
def _get_constructor_parameters(self) -> Dict[str, Any]:
data = super()._get_constructor_parameters()
data.update(
dict(
net_arch=self.net_arch,
features_dim=self.features_dim,
n_quantiles=self.n_quantiles,
activation_fn=self.activation_fn,
features_extractor=self.features_extractor,
)
)
return data
[docs]class QRDQNPolicy(BasePolicy):
"""
Policy class with quantile and target networks for QR-DQN.
:param observation_space: Observation space
:param action_space: Action space
:param lr_schedule: Learning rate schedule (could be constant)
:param n_quantiles: Number of quantiles
:param net_arch: The specification of the network architecture.
:param activation_fn: Activation function
:param features_extractor_class: Features extractor to use.
:param features_extractor_kwargs: Keyword arguments
to pass to the features extractor.
:param normalize_images: Whether to normalize images or not,
dividing by 255.0 (True by default)
:param optimizer_class: The optimizer to use,
``th.optim.Adam`` by default
:param optimizer_kwargs: Additional keyword arguments,
excluding the learning rate, to pass to the optimizer
"""
quantile_net: QuantileNetwork
quantile_net_target: QuantileNetwork
def __init__(
self,
observation_space: spaces.Space,
action_space: spaces.Discrete,
lr_schedule: Schedule,
n_quantiles: int = 200,
net_arch: Optional[List[int]] = None,
activation_fn: Type[nn.Module] = nn.ReLU,
features_extractor_class: Type[BaseFeaturesExtractor] = FlattenExtractor,
features_extractor_kwargs: Optional[Dict[str, Any]] = None,
normalize_images: bool = True,
optimizer_class: Type[th.optim.Optimizer] = th.optim.Adam,
optimizer_kwargs: Optional[Dict[str, Any]] = None,
):
super().__init__(
observation_space,
action_space,
features_extractor_class,
features_extractor_kwargs,
optimizer_class=optimizer_class,
optimizer_kwargs=optimizer_kwargs,
normalize_images=normalize_images,
)
if net_arch is None:
if features_extractor_class == NatureCNN:
net_arch = []
else:
net_arch = [64, 64]
self.n_quantiles = n_quantiles
self.net_arch = net_arch
self.activation_fn = activation_fn
self.net_args = {
"observation_space": self.observation_space,
"action_space": self.action_space,
"n_quantiles": self.n_quantiles,
"net_arch": self.net_arch,
"activation_fn": self.activation_fn,
"normalize_images": normalize_images,
}
self._build(lr_schedule)
def _build(self, lr_schedule: Schedule) -> None:
"""
Create the network and the optimizer.
:param lr_schedule: Learning rate schedule
lr_schedule(1) is the initial learning rate
"""
self.quantile_net = self.make_quantile_net()
self.quantile_net_target = self.make_quantile_net()
self.quantile_net_target.load_state_dict(self.quantile_net.state_dict())
self.quantile_net_target.set_training_mode(False)
# Setup optimizer with initial learning rate
self.optimizer = self.optimizer_class( # type: ignore[call-arg]
self.parameters(),
lr=lr_schedule(1),
**self.optimizer_kwargs,
)
def make_quantile_net(self) -> QuantileNetwork:
# Make sure we always have separate networks for features extractors etc
net_args = self._update_features_extractor(self.net_args, features_extractor=None)
return QuantileNetwork(**net_args).to(self.device)
[docs] def forward(self, obs: PyTorchObs, deterministic: bool = True) -> th.Tensor:
return self._predict(obs, deterministic=deterministic)
def _predict(self, obs: PyTorchObs, deterministic: bool = True) -> th.Tensor:
return self.quantile_net._predict(obs, deterministic=deterministic)
def _get_constructor_parameters(self) -> Dict[str, Any]:
data = super()._get_constructor_parameters()
data.update(
dict(
n_quantiles=self.net_args["n_quantiles"],
net_arch=self.net_args["net_arch"],
activation_fn=self.net_args["activation_fn"],
lr_schedule=self._dummy_schedule, # dummy lr schedule, not needed for loading policy alone
optimizer_class=self.optimizer_class,
optimizer_kwargs=self.optimizer_kwargs,
features_extractor_class=self.features_extractor_class,
features_extractor_kwargs=self.features_extractor_kwargs,
)
)
return data
[docs] def set_training_mode(self, mode: bool) -> None:
"""
Put the policy in either training or evaluation mode.
This affects certain modules, such as batch normalisation and dropout.
:param mode: if true, set to training mode, else set to evaluation mode
"""
self.quantile_net.set_training_mode(mode)
self.training = mode
MlpPolicy = QRDQNPolicy
[docs]class CnnPolicy(QRDQNPolicy):
"""
Policy class for QR-DQN when using images as input.
:param observation_space: Observation space
:param action_space: Action space
:param lr_schedule: Learning rate schedule (could be constant)
:param n_quantiles: Number of quantiles
:param net_arch: The specification of the network architecture.
:param activation_fn: Activation function
:param features_extractor_class: Features extractor to use.
:param normalize_images: Whether to normalize images or not,
dividing by 255.0 (True by default)
:param optimizer_class: The optimizer to use,
``th.optim.Adam`` by default
:param optimizer_kwargs: Additional keyword arguments,
excluding the learning rate, to pass to the optimizer
"""
def __init__(
self,
observation_space: spaces.Space,
action_space: spaces.Discrete,
lr_schedule: Schedule,
n_quantiles: int = 200,
net_arch: Optional[List[int]] = None,
activation_fn: Type[nn.Module] = nn.ReLU,
features_extractor_class: Type[BaseFeaturesExtractor] = NatureCNN,
features_extractor_kwargs: Optional[Dict[str, Any]] = None,
normalize_images: bool = True,
optimizer_class: Type[th.optim.Optimizer] = th.optim.Adam,
optimizer_kwargs: Optional[Dict[str, Any]] = None,
):
super().__init__(
observation_space,
action_space,
lr_schedule,
n_quantiles,
net_arch,
activation_fn,
features_extractor_class,
features_extractor_kwargs,
normalize_images,
optimizer_class,
optimizer_kwargs,
)