-
Notifications
You must be signed in to change notification settings - Fork 1
/
StaticSparseTD3.py
227 lines (167 loc) · 8.15 KB
/
StaticSparseTD3.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
227
import copy
import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
import sparse_utils as sp
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# Implementation of Twin Delayed Deep Deterministic Policy Gradients (TD3)
# Paper: https://arxiv.org/abs/1802.09477
class Actor(nn.Module):
def __init__(self, state_dim, action_dim, max_action,noHidNeurons,epsilonHid1,epsilonHid2):
super(Actor, self).__init__()
self.l1 = nn.Linear(state_dim, noHidNeurons)
[self.noPar1, self.mask1] = sp.initializeEpsilonWeightsMask("actor first layer", epsilonHid1, state_dim, noHidNeurons)
self.torchMask1=torch.from_numpy(self.mask1).float().to(device)
self.l1.weight.data.mul_(torch.from_numpy(self.mask1).float())
self.l2 = nn.Linear(noHidNeurons, noHidNeurons)
[self.noPar2, self.mask2] = sp.initializeEpsilonWeightsMask("actor second layer", epsilonHid2, noHidNeurons, noHidNeurons)
self.torchMask2 = torch.from_numpy(self.mask2).float().to(device)
self.l2.weight.data.mul_(torch.from_numpy(self.mask2).float())
self.l3 = nn.Linear(noHidNeurons, action_dim)
self.max_action = max_action
def forward(self, state):
a = F.relu(self.l1(state))
a = F.relu(self.l2(a))
return self.max_action * torch.tanh(self.l3(a))
class Critic(nn.Module):
def __init__(self, state_dim, action_dim,noHidNeurons,epsilonHid1,epsilonHid2):
super(Critic, self).__init__()
# Q1 architecture
self.l1 = nn.Linear(state_dim + action_dim, noHidNeurons)
[self.noPar1, self.mask1] = sp.initializeEpsilonWeightsMask("critic Q1 first layer", epsilonHid1, state_dim + action_dim, noHidNeurons)
self.torchMask1=torch.from_numpy(self.mask1).float().to(device)
self.l1.weight.data.mul_(torch.from_numpy(self.mask1).float())
self.l2 = nn.Linear(noHidNeurons, noHidNeurons)
[self.noPar2, self.mask2] = sp.initializeEpsilonWeightsMask("critic Q1 second layer", epsilonHid2, noHidNeurons, noHidNeurons)
self.torchMask2 = torch.from_numpy(self.mask2).float().to(device)
self.l2.weight.data.mul_(torch.from_numpy(self.mask2).float())
self.l3 = nn.Linear(noHidNeurons, 1)
# Q2 architecture
self.l4 = nn.Linear(state_dim + action_dim, noHidNeurons)
[self.noPar4, self.mask4] = sp.initializeEpsilonWeightsMask("critic Q2 first layer", epsilonHid1, state_dim + action_dim, noHidNeurons)
self.torchMask4 = torch.from_numpy(self.mask4).float().to(device)
self.l4.weight.data.mul_(torch.from_numpy(self.mask4).float())
self.l5 = nn.Linear(noHidNeurons, noHidNeurons)
[self.noPar5, self.mask5] = sp.initializeEpsilonWeightsMask("critic Q2 second layer", epsilonHid2, noHidNeurons, noHidNeurons)
self.torchMask5 = torch.from_numpy(self.mask5).float().to(device)
self.l5.weight.data.mul_(torch.from_numpy(self.mask5).float())
self.l6 = nn.Linear(noHidNeurons, 1)
def forward(self, state, action):
sa = torch.cat([state, action], 1)
q1 = F.relu(self.l1(sa))
q1 = F.relu(self.l2(q1))
q1 = self.l3(q1)
q2 = F.relu(self.l4(sa))
q2 = F.relu(self.l5(q2))
q2 = self.l6(q2)
return q1, q2
def Q1(self, state, action):
sa = torch.cat([state, action], 1)
q1 = F.relu(self.l1(sa))
q1 = F.relu(self.l2(q1))
q1 = self.l3(q1)
return q1
class StaticSparseTD3(object):
def __init__(
self,
state_dim,
action_dim,
max_action,
discount=0.99,
tau=0.005,
policy_noise=0.2,
noise_clip=0.5,
policy_freq=2,
noHidNeurons=256,
epsilonHid1=20,
epsilonHid2=20,
):
self.actor = Actor(state_dim, action_dim, max_action,noHidNeurons,epsilonHid1,epsilonHid2).to(device)
self.actor_target = copy.deepcopy(self.actor)
self.actor_optimizer = torch.optim.Adam(self.actor.parameters(), lr=1e-3, weight_decay=0.0002)
self.critic = Critic(state_dim, action_dim,noHidNeurons,epsilonHid1,epsilonHid2).to(device)
self.critic_target = copy.deepcopy(self.critic)
self.critic_optimizer = torch.optim.Adam(self.critic.parameters(), lr=1e-3, weight_decay=0.0002)
self.max_action = max_action
self.discount = discount
self.tau = tau
self.policy_noise = policy_noise
self.noise_clip = noise_clip
self.policy_freq = policy_freq
self.total_it = 0
def select_action(self, state):
state = torch.FloatTensor(state.reshape(1, -1)).to(device)
return self.actor(state).cpu().data.numpy().flatten()
def train(self, replay_buffer, batch_size=100):
self.total_it += 1
# Sample replay buffer
state, action, next_state, reward, not_done = replay_buffer.sample(batch_size)
with torch.no_grad():
# Select action according to policy and add clipped noise
noise = (
torch.randn_like(action) * self.policy_noise
).clamp(-self.noise_clip, self.noise_clip)
next_action = (
self.actor_target(next_state) + noise
).clamp(-self.max_action, self.max_action)
# Compute the target Q value
target_Q1, target_Q2 = self.critic_target(next_state, next_action)
target_Q = torch.min(target_Q1, target_Q2)
target_Q = reward + not_done * self.discount * target_Q
# Get current Q estimates
current_Q1, current_Q2 = self.critic(state, action)
# Compute critic loss
critic_loss = F.mse_loss(current_Q1, target_Q) + F.mse_loss(current_Q2, target_Q)
# Optimize the critic
self.critic_optimizer.zero_grad()
critic_loss.backward()
self.critic_optimizer.step()
# Maintain the same sparse connectivity for critic
self.critic.l1.weight.data.mul_(self.critic.torchMask1)
self.critic.l2.weight.data.mul_(self.critic.torchMask2)
self.critic.l4.weight.data.mul_(self.critic.torchMask4)
self.critic.l5.weight.data.mul_(self.critic.torchMask5)
# Delayed policy updates
if self.total_it % self.policy_freq == 0:
# Compute actor losse
actor_loss = -self.critic.Q1(state, self.actor(state)).mean()
# Optimize the actor
self.actor_optimizer.zero_grad()
actor_loss.backward()
self.actor_optimizer.step()
# Maintain the same sparse connectivity for actor
self.actor.l1.weight.data.mul_(self.actor.torchMask1)
self.actor.l2.weight.data.mul_(self.actor.torchMask2)
# Update the frozen target models
for param, target_param in zip(self.critic.parameters(), self.critic_target.parameters()):
target_param.data.copy_(self.tau * param.data + (1 - self.tau) * target_param.data)
for param, target_param in zip(self.actor.parameters(), self.actor_target.parameters()):
target_param.data.copy_(self.tau * param.data + (1 - self.tau) * target_param.data)
def print_sparsity(self):
for param, target_param in zip(self.critic.parameters(), self.critic_target.parameters()):
target_param.data.copy_(self.tau * param.data + (1 - self.tau) * target_param.data)
if(len(target_param.shape)>1):
critic_current_sparsity = ((target_param==0).sum().cpu().data.numpy()*1.0/(target_param.shape[0]*target_param.shape[1]))
print("target critic sparsity", critic_current_sparsity)
critic_current_sparsity = ((param==0).sum().cpu().data.numpy()*1.0/(param.shape[0]*param.shape[1]))
print("critic sparsity", critic_current_sparsity)
for param, target_param in zip(self.actor.parameters(), self.actor_target.parameters()):
target_param.data.copy_(self.tau * param.data + (1 - self.tau) * target_param.data)
if(len(target_param.shape)>1):
critic_current_sparsity = ((target_param==0).sum().cpu().data.numpy()*1.0/(target_param.shape[0]*target_param.shape[1]))
print("target actor sparsity", critic_current_sparsity)
critic_current_sparsity = ((param==0).sum().cpu().data.numpy()*1.0/(param.shape[0]*param.shape[1]))
print("actor sparsity", critic_current_sparsity)
def save(self, filename):
torch.save(self.critic.state_dict(), filename + "_critic")
torch.save(self.critic_optimizer.state_dict(), filename + "_critic_optimizer")
torch.save(self.actor.state_dict(), filename + "_actor")
torch.save(self.actor_optimizer.state_dict(), filename + "_actor_optimizer")
def load(self, filename):
self.critic.load_state_dict(torch.load(filename + "_critic"))
self.critic_optimizer.load_state_dict(torch.load(filename + "_critic_optimizer"))
self.critic_target = copy.deepcopy(self.critic)
self.actor.load_state_dict(torch.load(filename + "_actor"))
self.actor_optimizer.load_state_dict(torch.load(filename + "_actor_optimizer"))
self.actor_target = copy.deepcopy(self.actor)