from dataclasses import dataclass
from typing import Any, cast
import gymnasium as gym
import numpy as np
import torch
import torch.nn.functional as F
from overrides import override
from tianshou.algorithm import QRDQN, Algorithm
from tianshou.algorithm.modelfree.dqn import DiscreteQLearningPolicy
from tianshou.algorithm.modelfree.qrdqn import QRDQNPolicy
from tianshou.algorithm.modelfree.reinforce import SimpleLossTrainingStats
from tianshou.algorithm.optim import OptimizerFactory
from tianshou.data import Batch, ReplayBuffer, to_numpy
from tianshou.data.types import FQFBatchProtocol, ObsBatchProtocol, RolloutBatchProtocol
from tianshou.utils.net.discrete import FractionProposalNetwork, FullQuantileFunction
[docs]
@dataclass(kw_only=True)
class FQFTrainingStats(SimpleLossTrainingStats):
quantile_loss: float
fraction_loss: float
entropy_loss: float
[docs]
class FQFPolicy(QRDQNPolicy):
def __init__(
self,
*,
model: FullQuantileFunction,
fraction_model: FractionProposalNetwork,
action_space: gym.spaces.Space,
observation_space: gym.Space | None = None,
eps_training: float = 0.0,
eps_inference: float = 0.0,
):
"""
:param model: a model following the rules (s_B -> action_values_BA)
:param fraction_model: a FractionProposalNetwork for
proposing fractions/quantiles given state.
:param action_space: the environment's action space
:param observation_space: the environment's observation space.
:param eps_training: the epsilon value for epsilon-greedy exploration during training.
When collecting data for training, this is the probability of choosing a random action
instead of the action chosen by the policy.
A value of 0.0 means no exploration (fully greedy) and a value of 1.0 means full
exploration (fully random).
:param eps_inference: the epsilon value for epsilon-greedy exploration during inference,
i.e. non-training cases (such as evaluation during test steps).
The epsilon value is the probability of choosing a random action instead of the action
chosen by the policy.
A value of 0.0 means no exploration (fully greedy) and a value of 1.0 means full
exploration (fully random).
"""
assert isinstance(action_space, gym.spaces.Discrete)
super().__init__(
model=model,
action_space=action_space,
observation_space=observation_space,
eps_training=eps_training,
eps_inference=eps_inference,
)
self.fraction_model = fraction_model
[docs]
def forward( # type: ignore
self,
batch: ObsBatchProtocol,
state: dict | Batch | np.ndarray | None = None,
model: FullQuantileFunction | None = None,
fractions: Batch | None = None,
**kwargs: Any,
) -> FQFBatchProtocol:
if model is None:
model = self.model
obs = batch.obs
# TODO: this is convoluted! See also other places where this is done
obs_next = obs.obs if hasattr(obs, "obs") else obs
if fractions is None:
(logits, fractions, quantiles_tau), hidden = model(
obs_next,
propose_model=self.fraction_model,
state=state,
info=batch.info,
)
else:
(logits, _, quantiles_tau), hidden = model(
obs_next,
propose_model=self.fraction_model,
fractions=fractions,
state=state,
info=batch.info,
)
weighted_logits = (fractions.taus[:, 1:] - fractions.taus[:, :-1]).unsqueeze(1) * logits
q = DiscreteQLearningPolicy.compute_q_value(
self, weighted_logits.sum(2), getattr(obs, "mask", None)
)
act = to_numpy(q.max(dim=1)[1])
result = Batch(
logits=logits,
act=act,
state=hidden,
fractions=fractions,
quantiles_tau=quantiles_tau,
)
return cast(FQFBatchProtocol, result)
[docs]
class FQF(QRDQN[FQFPolicy]):
"""Implementation of Fully Parameterized Quantile Function for Distributional Reinforcement Learning. arXiv:1911.02140."""
def __init__(
self,
*,
policy: FQFPolicy,
optim: OptimizerFactory,
fraction_optim: OptimizerFactory,
gamma: float = 0.99,
num_fractions: int = 32,
ent_coef: float = 0.0,
n_step_return_horizon: int = 1,
target_update_freq: int = 0,
) -> None:
"""
:param policy: the policy
:param optim: the optimizer factory for the policy's main Q-function model
:param fraction_optim: the optimizer factory for the policy's fraction model
:param action_space: the environment's action space.
:param gamma: the discount factor in [0, 1] for future rewards.
This determines how much future rewards are valued compared to immediate ones.
Lower values (closer to 0) make the agent focus on immediate rewards, creating "myopic"
behavior. Higher values (closer to 1) make the agent value long-term rewards more,
potentially improving performance in tasks where delayed rewards are important but
increasing training variance by incorporating more environmental stochasticity.
Typically set between 0.9 and 0.99 for most reinforcement learning tasks
:param num_fractions: the number of fractions to use.
:param ent_coef: coefficient that weights the entropy bonus relative to the actor loss.
Controls the exploration-exploitation trade-off by encouraging policy entropy.
Higher values promote more exploration by encouraging a more uniform action distribution.
Lower values focus more on exploitation of the current policy's knowledge.
Typically set between 0.01 and 0.05 for most actor-critic implementations.
:param n_step_return_horizon: the number of future steps (> 0) to consider when computing temporal
difference (TD) targets. Controls the balance between TD learning and Monte Carlo methods:
higher values reduce bias (by relying less on potentially inaccurate value estimates)
but increase variance (by incorporating more environmental stochasticity and reducing
the averaging effect). A value of 1 corresponds to standard TD learning with immediate
bootstrapping, while very large values approach Monte Carlo-like estimation that uses
complete episode returns.
:param target_update_freq: the number of training iterations between each complete update of
the target network.
Controls how frequently the target Q-network parameters are updated with the current
Q-network values.
A value of 0 disables the target network entirely, using only a single network for both
action selection and bootstrap targets.
Higher values provide more stable learning targets but slow down the propagation of new
value estimates. Lower positive values allow faster learning but may lead to instability
due to rapidly changing targets.
Typically set between 100-10000 for DQN variants, with exact values depending on environment
complexity.
"""
super().__init__(
policy=policy,
optim=optim,
gamma=gamma,
num_quantiles=num_fractions,
n_step_return_horizon=n_step_return_horizon,
target_update_freq=target_update_freq,
)
self.ent_coef = ent_coef
self.fraction_optim = self._create_optimizer(self.policy.fraction_model, fraction_optim)
@override
def _create_policy_optimizer(self, optim: OptimizerFactory) -> Algorithm.Optimizer:
# Override to leave out the fraction model (use main model only), as we want
# to use a separate optimizer for the fraction model
return self._create_optimizer(self.policy.model, optim)
def _target_q(self, buffer: ReplayBuffer, indices: np.ndarray) -> torch.Tensor:
obs_next_batch = Batch(
obs=buffer[indices].obs_next,
info=[None] * len(indices),
) # obs_next: s_{t+n}
if self.use_target_network:
result = self.policy(obs_next_batch)
act, fractions = result.act, result.fractions
next_dist = self.policy(
obs_next_batch, model=self.model_old, fractions=fractions
).logits
else:
next_batch = self.policy(obs_next_batch)
act = next_batch.act
next_dist = next_batch.logits
return next_dist[np.arange(len(act)), act, :]
def _update_with_batch(
self,
batch: RolloutBatchProtocol,
) -> FQFTrainingStats:
self._periodically_update_lagged_network_weights()
weight = batch.pop("weight", 1.0)
out = self.policy(batch)
curr_dist_orig = out.logits
taus, tau_hats = out.fractions.taus, out.fractions.tau_hats
act = batch.act
curr_dist = curr_dist_orig[np.arange(len(act)), act, :].unsqueeze(2)
target_dist = batch.returns.unsqueeze(1)
# calculate each element's difference between curr_dist and target_dist
dist_diff = F.smooth_l1_loss(target_dist, curr_dist, reduction="none")
huber_loss = (
(
dist_diff
* (tau_hats.unsqueeze(2) - (target_dist - curr_dist).detach().le(0.0).float()).abs()
)
.sum(-1)
.mean(1)
)
quantile_loss = (huber_loss * weight).mean()
# ref: https://github.com/ku2482/fqf-iqn-qrdqn.pytorch/
# blob/master/fqf_iqn_qrdqn/agent/qrdqn_agent.py L130
batch.weight = dist_diff.detach().abs().sum(-1).mean(1) # prio-buffer
# calculate fraction loss
with torch.no_grad():
sa_quantile_hats = curr_dist_orig[np.arange(len(act)), act, :]
sa_quantiles = out.quantiles_tau[np.arange(len(act)), act, :]
# ref: https://github.com/ku2482/fqf-iqn-qrdqn.pytorch/
# blob/master/fqf_iqn_qrdqn/agent/fqf_agent.py L169
values_1 = sa_quantiles - sa_quantile_hats[:, :-1]
signs_1 = sa_quantiles > torch.cat(
[sa_quantile_hats[:, :1], sa_quantiles[:, :-1]],
dim=1,
)
values_2 = sa_quantiles - sa_quantile_hats[:, 1:]
signs_2 = sa_quantiles < torch.cat(
[sa_quantiles[:, 1:], sa_quantile_hats[:, -1:]],
dim=1,
)
gradient_of_taus = torch.where(signs_1, values_1, -values_1) + torch.where(
signs_2,
values_2,
-values_2,
)
fraction_loss = (gradient_of_taus * taus[:, 1:-1]).sum(1).mean()
# calculate entropy loss
entropy_loss = out.fractions.entropies.mean()
fraction_entropy_loss = fraction_loss - self.ent_coef * entropy_loss
self.fraction_optim.step(fraction_entropy_loss, retain_graph=True)
self.optim.step(quantile_loss)
return FQFTrainingStats(
loss=quantile_loss.item() + fraction_entropy_loss.item(),
quantile_loss=quantile_loss.item(),
fraction_loss=fraction_loss.item(),
entropy_loss=entropy_loss.item(),
)
