File size: 7,302 Bytes
8e63d2a
 
 
79943a9
 
 
48f24b9
79943a9
48f24b9
 
 
 
 
 
79943a9
 
 
 
8e63d2a
79943a9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8e63d2a
 
 
 
 
 
79943a9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8e63d2a
79943a9
 
 
8e63d2a
79943a9
 
 
 
 
 
48f24b9
79943a9
 
 
 
48f24b9
 
79943a9
48f24b9
 
 
79943a9
48f24b9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8e63d2a
79943a9
 
48f24b9
 
 
79943a9
 
48f24b9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
79943a9
8e63d2a
 
 
 
 
 
48f24b9
79943a9
 
 
 
 
 
 
 
48f24b9
79943a9
 
8e63d2a
79943a9
8e63d2a
79943a9
48f24b9
79943a9
 
 
 
 
 
 
8e63d2a
 
 
 
 
 
79943a9
8e63d2a
79943a9
 
8e63d2a
 
 
79943a9
48f24b9
 
 
79943a9
 
 
 
 
 
 
48f24b9
 
 
 
 
79943a9
48f24b9
8e63d2a
 
 
 
 
48f24b9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
from abc import abstractmethod
from typing import NamedTuple, Optional, Sequence, Tuple, TypeVar

import gym
import numpy as np
import torch
from gym.spaces import Box, Space

from rl_algo_impls.shared.policy.actor_critic_network import (
    ConnectedTrioActorCriticNetwork,
    SeparateActorCriticNetwork,
    UNetActorCriticNetwork,
)
from rl_algo_impls.shared.policy.policy import Policy
from rl_algo_impls.wrappers.vectorable_wrapper import (
    VecEnv,
    VecEnvObs,
    single_action_space,
    single_observation_space,
)


class Step(NamedTuple):
    a: np.ndarray
    v: np.ndarray
    logp_a: np.ndarray
    clamped_a: np.ndarray


class ACForward(NamedTuple):
    logp_a: torch.Tensor
    entropy: torch.Tensor
    v: torch.Tensor


FEAT_EXT_FILE_NAME = "feat_ext.pt"
V_FEAT_EXT_FILE_NAME = "v_feat_ext.pt"
PI_FILE_NAME = "pi.pt"
V_FILE_NAME = "v.pt"
ActorCriticSelf = TypeVar("ActorCriticSelf", bound="ActorCritic")


def clamp_actions(
    actions: np.ndarray, action_space: gym.Space, squash_output: bool
) -> np.ndarray:
    if isinstance(action_space, Box):
        low, high = action_space.low, action_space.high  # type: ignore
        if squash_output:
            # Squashed output is already between -1 and 1. Rescale if the actual
            # output needs to something other than -1 and 1
            return low + 0.5 * (actions + 1) * (high - low)
        else:
            return np.clip(actions, low, high)
    return actions


class OnPolicy(Policy):
    @abstractmethod
    def value(self, obs: VecEnvObs) -> np.ndarray:
        ...

    @abstractmethod
    def step(self, obs: VecEnvObs, action_masks: Optional[np.ndarray] = None) -> Step:
        ...

    @property
    @abstractmethod
    def action_shape(self) -> Tuple[int, ...]:
        ...


class ActorCritic(OnPolicy):
    def __init__(
        self,
        env: VecEnv,
        pi_hidden_sizes: Optional[Sequence[int]] = None,
        v_hidden_sizes: Optional[Sequence[int]] = None,
        init_layers_orthogonal: bool = True,
        activation_fn: str = "tanh",
        log_std_init: float = -0.5,
        use_sde: bool = False,
        full_std: bool = True,
        squash_output: bool = False,
        share_features_extractor: bool = True,
        cnn_flatten_dim: int = 512,
        cnn_style: str = "nature",
        cnn_layers_init_orthogonal: Optional[bool] = None,
        impala_channels: Sequence[int] = (16, 32, 32),
        actor_head_style: str = "single",
        **kwargs,
    ) -> None:
        super().__init__(env, **kwargs)

        observation_space = single_observation_space(env)
        action_space = single_action_space(env)
        action_plane_space = getattr(env, "action_plane_space", None)

        self.action_space = action_space
        self.squash_output = squash_output

        if actor_head_style == "unet":
            self.network = UNetActorCriticNetwork(
                observation_space,
                action_space,
                action_plane_space,
                v_hidden_sizes=v_hidden_sizes,
                init_layers_orthogonal=init_layers_orthogonal,
                activation_fn=activation_fn,
                cnn_layers_init_orthogonal=cnn_layers_init_orthogonal,
            )
        elif share_features_extractor:
            self.network = ConnectedTrioActorCriticNetwork(
                observation_space,
                action_space,
                pi_hidden_sizes=pi_hidden_sizes,
                v_hidden_sizes=v_hidden_sizes,
                init_layers_orthogonal=init_layers_orthogonal,
                activation_fn=activation_fn,
                log_std_init=log_std_init,
                use_sde=use_sde,
                full_std=full_std,
                squash_output=squash_output,
                cnn_flatten_dim=cnn_flatten_dim,
                cnn_style=cnn_style,
                cnn_layers_init_orthogonal=cnn_layers_init_orthogonal,
                impala_channels=impala_channels,
                actor_head_style=actor_head_style,
                action_plane_space=action_plane_space,
            )
        else:
            self.network = SeparateActorCriticNetwork(
                observation_space,
                action_space,
                pi_hidden_sizes=pi_hidden_sizes,
                v_hidden_sizes=v_hidden_sizes,
                init_layers_orthogonal=init_layers_orthogonal,
                activation_fn=activation_fn,
                log_std_init=log_std_init,
                use_sde=use_sde,
                full_std=full_std,
                squash_output=squash_output,
                cnn_flatten_dim=cnn_flatten_dim,
                cnn_style=cnn_style,
                cnn_layers_init_orthogonal=cnn_layers_init_orthogonal,
                impala_channels=impala_channels,
                actor_head_style=actor_head_style,
                action_plane_space=action_plane_space,
            )

    def forward(
        self,
        obs: torch.Tensor,
        action: torch.Tensor,
        action_masks: Optional[torch.Tensor] = None,
    ) -> ACForward:
        (_, logp_a, entropy), v = self.network(obs, action, action_masks=action_masks)

        assert logp_a is not None
        assert entropy is not None
        return ACForward(logp_a, entropy, v)

    def value(self, obs: VecEnvObs) -> np.ndarray:
        o = self._as_tensor(obs)
        with torch.no_grad():
            v = self.network.value(o)
        return v.cpu().numpy()

    def step(self, obs: VecEnvObs, action_masks: Optional[np.ndarray] = None) -> Step:
        o = self._as_tensor(obs)
        a_masks = self._as_tensor(action_masks) if action_masks is not None else None
        with torch.no_grad():
            (pi, _, _), v = self.network.distribution_and_value(o, action_masks=a_masks)
            a = pi.sample()
            logp_a = pi.log_prob(a)

        a_np = a.cpu().numpy()
        clamped_a_np = clamp_actions(a_np, self.action_space, self.squash_output)
        return Step(a_np, v.cpu().numpy(), logp_a.cpu().numpy(), clamped_a_np)

    def act(
        self,
        obs: np.ndarray,
        deterministic: bool = True,
        action_masks: Optional[np.ndarray] = None,
    ) -> np.ndarray:
        if not deterministic:
            return self.step(obs, action_masks=action_masks).clamped_a
        else:
            o = self._as_tensor(obs)
            a_masks = (
                self._as_tensor(action_masks) if action_masks is not None else None
            )
            with torch.no_grad():
                (pi, _, _), _ = self.network.distribution_and_value(
                    o, action_masks=a_masks
                )
                a = pi.mode
            return clamp_actions(a.cpu().numpy(), self.action_space, self.squash_output)

    def load(self, path: str) -> None:
        super().load(path)
        self.reset_noise()

    def load_from(self: ActorCriticSelf, policy: ActorCriticSelf) -> ActorCriticSelf:
        super().load_from(policy)
        self.reset_noise()
        return self

    def reset_noise(self, batch_size: Optional[int] = None) -> None:
        self.network.reset_noise(
            batch_size=batch_size if batch_size else self.env.num_envs
        )

    @property
    def action_shape(self) -> Tuple[int, ...]:
        return self.network.action_shape