-
-
Notifications
You must be signed in to change notification settings - Fork 953
/
Copy pathfunctional_jax_env.py
226 lines (167 loc) · 7.14 KB
/
functional_jax_env.py
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
218
219
220
221
222
223
224
225
226
"""Functional to Environment compatibility."""
from __future__ import annotations
from typing import Any
import jax
import jax.numpy as jnp
import jax.random as jrng
import gymnasium as gym
from gymnasium.envs.registration import EnvSpec
from gymnasium.experimental.functional import ActType, FuncEnv, StateType
from gymnasium.utils import seeding
from gymnasium.vector.utils import batch_space
class FunctionalJaxEnv(gym.Env):
"""A conversion layer for jax-based environments."""
state: StateType
rng: jrng.PRNGKey
def __init__(
self,
func_env: FuncEnv,
metadata: dict[str, Any] | None = None,
render_mode: str | None = None,
spec: EnvSpec | None = None,
):
"""Initialize the environment from a FuncEnv."""
if metadata is None:
# metadata.get("jax", False) can be used downstream to know that the environment returns jax arrays
metadata = {"render_mode": [], "jax": True}
self.func_env = func_env
self.observation_space = func_env.observation_space
self.action_space = func_env.action_space
self.metadata = metadata
self.render_mode = render_mode
self.spec = spec
if self.render_mode == "rgb_array":
self.render_state = self.func_env.render_init()
else:
self.render_state = None
np_random, _ = seeding.np_random()
seed = np_random.integers(0, 2**32 - 1, dtype="uint32")
self.rng = jrng.PRNGKey(seed)
def reset(self, *, seed: int | None = None, options: dict | None = None):
"""Resets the environment using the seed."""
super().reset(seed=seed)
if seed is not None:
self.rng = jrng.PRNGKey(seed)
rng, self.rng = jrng.split(self.rng)
self.state = self.func_env.initial(rng=rng)
obs = self.func_env.observation(self.state, rng)
info = self.func_env.state_info(self.state)
return obs, info
def step(self, action: ActType):
"""Steps through the environment using the action."""
rng, self.rng = jrng.split(self.rng)
next_state = self.func_env.transition(self.state, action, rng)
observation = self.func_env.observation(next_state, rng)
reward = self.func_env.reward(self.state, action, next_state, rng)
terminated = self.func_env.terminal(next_state, rng)
info = self.func_env.transition_info(self.state, action, next_state)
self.state = next_state
return observation, float(reward), bool(terminated), False, info
def render(self):
"""Returns the render state if `render_mode` is "rgb_array"."""
if self.render_mode == "rgb_array":
self.render_state, image = self.func_env.render_image(
self.state, self.render_state
)
return image
else:
raise NotImplementedError
def close(self):
"""Closes the environments and render state if set."""
if self.render_state is not None:
self.func_env.render_close(self.render_state)
self.render_state = None
class FunctionalJaxVectorEnv(gym.vector.VectorEnv):
"""A vector env implementation for functional Jax envs."""
state: StateType
rng: jrng.PRNGKey
def __init__(
self,
func_env: FuncEnv,
num_envs: int,
max_episode_steps: int = 0,
metadata: dict[str, Any] | None = None,
render_mode: str | None = None,
spec: EnvSpec | None = None,
):
"""Initialize the environment from a FuncEnv."""
super().__init__()
if metadata is None:
metadata = {}
self.func_env = func_env
self.num_envs = num_envs
self.single_observation_space = func_env.observation_space
self.single_action_space = func_env.action_space
self.observation_space = batch_space(
self.single_observation_space, self.num_envs
)
self.action_space = batch_space(self.single_action_space, self.num_envs)
self.metadata = metadata
self.render_mode = render_mode
self.spec = spec
self.time_limit = max_episode_steps
self.steps = jnp.zeros(self.num_envs, dtype=jnp.int32)
self.prev_done = jnp.zeros(self.num_envs, dtype=jnp.bool_)
if self.render_mode == "rgb_array":
self.render_state = self.func_env.render_init()
else:
self.render_state = None
np_random, _ = seeding.np_random()
seed = np_random.integers(0, 2**32 - 1, dtype="uint32")
self.rng = jrng.PRNGKey(seed)
self.func_env.transform(jax.vmap)
def reset(self, *, seed: int | None = None, options: dict | None = None):
"""Resets the environment."""
super().reset(seed=seed)
if seed is not None:
self.rng = jrng.PRNGKey(seed)
rng, self.rng = jrng.split(self.rng)
rng = jrng.split(rng, self.num_envs)
self.state = self.func_env.initial(rng=rng)
obs = self.func_env.observation(self.state, rng)
info = self.func_env.state_info(self.state)
self.steps = jnp.zeros(self.num_envs, dtype=jnp.int32)
return obs, info
def step(self, action: ActType):
"""Steps through the environment using the action."""
self.steps += 1
rng, self.rng = jrng.split(self.rng)
rng = jrng.split(rng, self.num_envs)
next_state = self.func_env.transition(self.state, action, rng)
reward = self.func_env.reward(self.state, action, next_state, rng)
terminated = self.func_env.terminal(next_state, rng)
truncated = (
self.steps >= self.time_limit
if self.time_limit > 0
else jnp.zeros_like(terminated)
)
info = self.func_env.transition_info(self.state, action, next_state)
if jnp.any(self.prev_done):
to_reset = jnp.where(self.prev_done)[0]
reset_count = to_reset.shape[0]
rng, self.rng = jrng.split(self.rng)
rng = jrng.split(rng, reset_count)
new_initials = self.func_env.initial(rng)
next_state = self.state.at[to_reset].set(new_initials)
self.steps = self.steps.at[to_reset].set(0)
terminated = terminated.at[to_reset].set(False)
truncated = truncated.at[to_reset].set(False)
self.prev_done = jnp.logical_or(terminated, truncated)
rng = jrng.split(self.rng, self.num_envs)
observation = self.func_env.observation(next_state, rng)
self.state = next_state
return observation, reward, terminated, truncated, info
def render(self):
"""Returns the render state if `render_mode` is "rgb_array"."""
if self.render_mode == "rgb_array":
self.render_state, image = self.func_env.render_image(
self.state, self.render_state
)
return image
else:
raise NotImplementedError
def close(self):
"""Closes the environments and render state if set."""
if self.render_state is not None:
self.func_env.render_close(self.render_state)
self.render_state = None