您最多选择25个主题
主题必须以中文或者字母或数字开头，可以包含连字符 (-)，并且长度不得超过35个字符
97 行
4.1 KiB
97 行
4.1 KiB
from typing import Dict, Optional, Tuple, List
|
|
from mlagents.torch_utils import torch
|
|
import numpy as np
|
|
|
|
from mlagents.trainers.buffer import AgentBuffer
|
|
from mlagents.trainers.trajectory import SplitObservations
|
|
from mlagents.trainers.torch.components.bc.module import BCModule
|
|
from mlagents.trainers.torch.components.reward_providers import create_reward_provider
|
|
|
|
from mlagents.trainers.policy.torch_policy import TorchPolicy
|
|
from mlagents.trainers.optimizer import Optimizer
|
|
from mlagents.trainers.settings import TrainerSettings
|
|
from mlagents.trainers.torch.utils import ModelUtils
|
|
|
|
|
|
class TorchOptimizer(Optimizer): # pylint: disable=W0223
|
|
def __init__(self, policy: TorchPolicy, trainer_settings: TrainerSettings):
|
|
super().__init__()
|
|
self.policy = policy
|
|
self.trainer_settings = trainer_settings
|
|
self.update_dict: Dict[str, torch.Tensor] = {}
|
|
self.value_heads: Dict[str, torch.Tensor] = {}
|
|
self.memory_in: torch.Tensor = None
|
|
self.memory_out: torch.Tensor = None
|
|
self.m_size: int = 0
|
|
self.global_step = torch.tensor(0)
|
|
self.bc_module: Optional[BCModule] = None
|
|
self.create_reward_signals(trainer_settings.reward_signals)
|
|
if trainer_settings.behavioral_cloning is not None:
|
|
self.bc_module = BCModule(
|
|
self.policy,
|
|
trainer_settings.behavioral_cloning,
|
|
policy_learning_rate=trainer_settings.hyperparameters.learning_rate,
|
|
default_batch_size=trainer_settings.hyperparameters.batch_size,
|
|
default_num_epoch=3,
|
|
)
|
|
|
|
def update(self, batch: AgentBuffer, num_sequences: int) -> Dict[str, float]:
|
|
pass
|
|
|
|
def create_reward_signals(self, reward_signal_configs):
|
|
"""
|
|
Create reward signals
|
|
:param reward_signal_configs: Reward signal config.
|
|
"""
|
|
for reward_signal, settings in reward_signal_configs.items():
|
|
# Name reward signals by string in case we have duplicates later
|
|
self.reward_signals[reward_signal.value] = create_reward_provider(
|
|
reward_signal, self.policy.behavior_spec, settings
|
|
)
|
|
|
|
def get_trajectory_value_estimates(
|
|
self, batch: AgentBuffer, next_obs: List[np.ndarray], done: bool
|
|
) -> Tuple[Dict[str, np.ndarray], Dict[str, float]]:
|
|
vector_obs = [ModelUtils.list_to_tensor(batch["vector_obs"])[:, 1:]]
|
|
goals = [ModelUtils.list_to_tensor(batch["vector_obs"])[:, :1]]
|
|
if self.policy.use_vis_obs:
|
|
visual_obs = []
|
|
for idx, _ in enumerate(
|
|
self.policy.actor_critic.network_body.visual_processors
|
|
):
|
|
visual_ob = ModelUtils.list_to_tensor(batch["visual_obs%d" % idx])
|
|
visual_obs.append(visual_ob)
|
|
else:
|
|
visual_obs = []
|
|
|
|
memory = torch.zeros([1, 1, self.policy.m_size])
|
|
|
|
vec_vis_obs = SplitObservations.from_observations(next_obs)
|
|
next_vec_obs = [
|
|
ModelUtils.list_to_tensor(vec_vis_obs.vector_observations[1:]).unsqueeze(0)
|
|
]
|
|
next_vis_obs = [
|
|
ModelUtils.list_to_tensor(_vis_ob).unsqueeze(0)
|
|
for _vis_ob in vec_vis_obs.visual_observations
|
|
]
|
|
|
|
# goals dont change but otherwise broken
|
|
next_goals = [torch.as_tensor(vec_vis_obs.vector_observations[:1])]
|
|
value_estimates, next_memory = self.policy.actor_critic.critic_pass(
|
|
vector_obs, visual_obs, goals, memory, sequence_length=batch.num_experiences
|
|
)
|
|
|
|
next_value_estimate, _ = self.policy.actor_critic.critic_pass(
|
|
next_vec_obs, next_vis_obs, next_goals, next_memory, sequence_length=1
|
|
)
|
|
|
|
for name, estimate in value_estimates.items():
|
|
value_estimates[name] = ModelUtils.to_numpy(estimate)
|
|
next_value_estimate[name] = ModelUtils.to_numpy(next_value_estimate[name])
|
|
|
|
if done:
|
|
for k in next_value_estimate:
|
|
if not self.reward_signals[k].ignore_done:
|
|
next_value_estimate[k] = 0.0
|
|
|
|
return value_estimates, next_value_estimate
|