PARL/examples/PPO/storage.py at develop · lyyongdev/PARL · GitHub

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
#   Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import numpy as np


class RolloutStorage():
    def __init__(self, step_nums, env_num, obs_space, act_space):
        self.obs = np.zeros(
            (step_nums, env_num) + obs_space.shape, dtype='float32')
        self.actions = np.zeros(
            (step_nums, env_num) + act_space.shape, dtype='float32')
        self.logprobs = np.zeros((step_nums, env_num), dtype='float32')
        self.rewards = np.zeros((step_nums, env_num), dtype='float32')
        self.dones = np.zeros((step_nums, env_num), dtype='float32')
        self.values = np.zeros((step_nums, env_num), dtype='float32')

        self.step_nums = step_nums
        self.obs_space = obs_space
        self.act_space = act_space

        self.cur_step = 0

    def append(self, obs, action, logprob, reward, done, value):
        self.obs[self.cur_step] = obs
        self.actions[self.cur_step] = action
        self.logprobs[self.cur_step] = logprob
        self.rewards[self.cur_step] = reward
        self.dones[self.cur_step] = done
        self.values[self.cur_step] = value

        self.cur_step = (self.cur_step + 1) % self.step_nums

    def compute_returns(self, value, done, gamma=0.99, gae_lambda=0.95):
        # gamma: discounting factor
        # gae_lambda: Lambda parameter for calculating N-step advantage
        advantages = np.zeros_like(self.rewards)
        lastgaelam = 0
        for t in reversed(range(self.step_nums)):
            if t == self.step_nums - 1:
                nextnonterminal = 1.0 - done
                nextvalues = value.reshape(1, -1)
            else:
                nextnonterminal = 1.0 - self.dones[t + 1]
                nextvalues = self.values[t + 1]
            delta = self.rewards[
                t] + gamma * nextvalues * nextnonterminal - self.values[t]
            advantages[
                t] = lastgaelam = delta + gamma * gae_lambda * nextnonterminal * lastgaelam
        returns = advantages + self.values
        self.returns = returns
        self.advantages = advantages
        return advantages, returns

    def sample_batch(self, idx):
        # flatten rollout
        b_obs = self.obs.reshape((-1, ) + self.obs_space.shape)
        b_logprobs = self.logprobs.reshape(-1)
        b_actions = self.actions.reshape((-1, ) + self.act_space.shape)
        b_advantages = self.advantages.reshape(-1)
        b_returns = self.returns.reshape(-1)
        b_values = self.values.reshape(-1)

        return b_obs[idx], b_actions[idx], b_logprobs[idx], b_advantages[
            idx], b_returns[idx], b_values[idx]