-
Notifications
You must be signed in to change notification settings - Fork 0
/
gail.py
379 lines (315 loc) · 15.4 KB
/
gail.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
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
import argparse
import sys
import math
from collections import namedtuple
from itertools import count
import gym
import numpy as np
import scipy.optimize
from gym import wrappers
import torch
import torch.autograd as autograd
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
import torchvision.transforms as T
from torch.autograd import Variable
from models import Policy, Value, Reward, ActorCritic
from replay_memory import Memory
from load_expert_traj import Expert
from running_state import ZFilter
# from utils import *
torch.set_default_tensor_type('torch.DoubleTensor')
PI = torch.DoubleTensor([3.1415926])
parser = argparse.ArgumentParser(description='PyTorch actor-critic example')
parser.add_argument('--gamma', type=float, default=0.99, metavar='G',
help='discount factor (default: 0.99)')
parser.add_argument('--env-name', default="Hopper-v1", metavar='G',
help='name of the environment to run')
parser.add_argument('--expert-path', default="hopper_expert_trajectories/", metavar='G',
help='path to the expert trajectory files')
parser.add_argument('--tau', type=float, default=0.95, metavar='G',
help='gae (default: 0.95)')
parser.add_argument('--learning-rate', type=float, default=3e-4, metavar='G',
help='gae (default: 3e-4)')
parser.add_argument('--seed', type=int, default=1, metavar='N',
help='random seed (default: 1)')
parser.add_argument('--batch-size', type=int, default=2048, metavar='N',
help='batch size (default: 2048)')
parser.add_argument('--num-episodes', type=int, default=500, metavar='N',
help='number of episodes (default: 500)')
parser.add_argument('--optim-epochs', type=int, default=5, metavar='N',
help='number of epochs over a batch (default: 5)')
parser.add_argument('--optim-batch-size', type=int, default=64, metavar='N',
help='batch size for epochs (default: 64)')
parser.add_argument('--num-expert-trajs', type=int, default=5, metavar='N',
help='number of expert trajectories in a batch (default: 5)')
parser.add_argument('--render', action='store_true',
help='render the environment')
parser.add_argument('--log-interval', type=int, default=1, metavar='N',
help='interval between training status logs (default: 10)')
parser.add_argument('--save-interval', type=int, default=100, metavar='N',
help='interval between saving policy weights (default: 100)')
parser.add_argument('--entropy-coeff', type=float, default=0.0, metavar='N',
help='coefficient for entropy cost')
parser.add_argument('--clip-epsilon', type=float, default=0.2, metavar='N',
help='Clipping for PPO grad')
parser.add_argument('--use-joint-pol-val', action='store_true',
help='whether to use combined policy and value nets')
args = parser.parse_args()
env = gym.make(args.env_name)
num_inputs = env.observation_space.shape[0]
num_actions = env.action_space.shape[0]
env.seed(args.seed)
torch.manual_seed(args.seed)
if args.use_joint_pol_val:
ac_net = ActorCritic(num_inputs, num_actions)
opt_ac = optim.Adam(ac_net.parameters(), lr=0.0003)
else:
policy_net = Policy(num_inputs, num_actions)
old_policy_net = Policy(num_inputs, num_actions)
value_net = Value(num_inputs)
reward_net = Reward(num_inputs, num_actions)
opt_policy = optim.Adam(policy_net.parameters(), lr=0.0003)
opt_value = optim.Adam(value_net.parameters(), lr=0.0003)
opt_reward = optim.Adam(reward_net.parameters(), lr=0.0003)
def select_action(state):
state = torch.from_numpy(state).unsqueeze(0)
action_mean, _, action_std = policy_net(Variable(state))
action = torch.normal(action_mean, action_std)
return action
def select_action_actor_critic(state):
state = torch.from_numpy(state).unsqueeze(0)
action_mean, _, action_std, v = ac_net(Variable(state))
action = torch.normal(action_mean, action_std)
return action
def normal_log_density(x, mean, log_std, std):
var = std.pow(2)
log_density = -(x - mean).pow(2) / (2 * var) - 0.5 * torch.log(2 * Variable(PI)) - log_std
return log_density.sum(1)
def update_params_actor_critic(batch, i_episode):
rewards = torch.Tensor(batch.reward)
masks = torch.Tensor(batch.mask)
actions = torch.Tensor(np.concatenate(batch.action, 0))
states = torch.Tensor(batch.state)
action_means, action_log_stds, action_stds, values = ac_net(Variable(states))
returns = torch.Tensor(actions.size(0),1)
deltas = torch.Tensor(actions.size(0),1)
advantages = torch.Tensor(actions.size(0),1)
opt_ac.lr = args.learning_rate*max(1.0 - float(i_episode)/args.num_episodes, 0)
clip_epsilon = args.clip_epsilon*max(1.0 - float(i_episode)/args.num_episodes, 0)
prev_return = 0
prev_value = 0
prev_advantage = 0
for i in reversed(range(rewards.size(0))):
returns[i] = rewards[i] + args.gamma * prev_return * masks[i]
deltas[i] = rewards[i] + args.gamma * prev_value * masks[i] - values.data[i]
advantages[i] = deltas[i] + args.gamma * args.tau * prev_advantage * masks[i]
prev_return = returns[i, 0]
prev_value = values.data[i, 0]
prev_advantage = advantages[i, 0]
targets = Variable(returns)
# kloldnew = policy_net.kl_old_new() # oldpi.pd.kl(pi.pd)
# ent = policy_net.entropy() #pi.pd.entropy()
# meankl = torch.reduce_mean(kloldnew)
# meanent = torch.reduce_mean(ent)
# pol_entpen = (-args.entropy_coeff) * meanent
action_var = Variable(actions)
# compute probs from actions above
log_prob_cur = normal_log_density(action_var, action_means, action_log_stds, action_stds)
action_means_old, action_log_stds_old, action_stds_old, values_old = ac_net(Variable(states), old=True)
log_prob_old = normal_log_density(action_var, action_means_old, action_log_stds_old, action_stds_old)
# backup params after computing probs but before updating new params
ac_net.backup()
advantages = (advantages - advantages.mean()) / advantages.std()
advantages_var = Variable(advantages)
opt_ac.zero_grad()
ratio = torch.exp(log_prob_cur - log_prob_old) # pnew / pold
surr1 = ratio * advantages_var[:,0]
surr2 = torch.clamp(ratio, 1.0 - clip_epsilon, 1.0 + clip_epsilon) * advantages_var[:,0]
policy_surr = -torch.min(surr1, surr2).mean()
vf_loss = (values - targets).pow(2.).mean()
#vpredclipped = values_old + torch.clamp(values - values_old, -args.clip_epsilon, args.clip_epsilon)
#vf_loss2 = (vpredclipped - targets).pow(2.)
#vf_loss = 0.5 * torch.max(vf_loss1, vf_loss2).mean()
total_loss = policy_surr + vf_loss
total_loss.backward()
#torch.nn.utils.clip_grad_norm(ac_net.parameters(), 40)
opt_ac.step()
def update_params(gen_batch, expert_batch, i_episode, optim_epochs, optim_batch_size):
criterion = nn.BCELoss()
# generated trajectories
rewards = torch.Tensor(gen_batch.reward)
masks = torch.Tensor(gen_batch.mask)
actions = torch.Tensor(np.concatenate(gen_batch.action, 0))
states = torch.Tensor(gen_batch.state)
values = value_net(Variable(states))
# expert trajectories
list_of_expert_states = []
for i in range(len(expert_batch.state)):
list_of_expert_states.append(torch.Tensor(expert_batch.state[i]))
expert_states = torch.cat(list_of_expert_states,0)
list_of_expert_actions = []
for i in range(len(expert_batch.action)):
list_of_expert_actions.append(torch.Tensor(expert_batch.action[i]))
expert_actions = torch.cat(list_of_expert_actions, 0)
list_of_masks = []
for i in range(len(expert_batch.mask)):
list_of_masks.append(torch.Tensor(expert_batch.mask[i]))
expert_masks = torch.cat(list_of_masks, 0)
returns = torch.Tensor(actions.size(0),1)
deltas = torch.Tensor(actions.size(0),1)
advantages = torch.Tensor(actions.size(0),1)
opt_value.lr = args.learning_rate*max(1.0 - float(i_episode)/args.num_episodes, 0)
opt_policy.lr = args.learning_rate*max(1.0 - float(i_episode)/args.num_episodes, 0)
opt_reward.lr = args.learning_rate*max(1.0 - float(i_episode)/args.num_episodes, 0)
clip_epsilon = args.clip_epsilon*max(1.0 - float(i_episode)/args.num_episodes, 0)
# compute advantages
prev_return = 0
prev_value = 0
prev_advantage = 0
for i in reversed(range(rewards.size(0))):
returns[i] = rewards[i] + args.gamma * prev_return * masks[i]
deltas[i] = rewards[i] + args.gamma * prev_value * masks[i] - values.data[i]
advantages[i] = deltas[i] + args.gamma * args.tau * prev_advantage * masks[i]
prev_return = returns[i, 0]
prev_value = values.data[i, 0]
prev_advantage = advantages[i, 0]
targets = Variable(returns)
advantages = (advantages - advantages.mean()) / advantages.std()
# backup params after computing probs but before updating new params
#policy_net.backup()
for old_policy_param, policy_param in zip(old_policy_net.parameters(), policy_net.parameters()):
old_policy_param.data.copy_(policy_param.data)
# kloldnew = policy_net.kl_old_new() # oldpi.pd.kl(pi.pd)
# ent = policy_net.entropy() #pi.pd.entropy()
# meankl = torch.reduce_mean(kloldnew)
# meanent = torch.reduce_mean(ent)
# pol_entpen = (-args.entropy_coeff) * meanent
# update value, reward and policy networks
optim_iters = int(math.ceil(args.batch_size/optim_batch_size))
optim_batch_size_exp = int(math.floor(expert_actions.size(0)/(optim_iters)))
for _ in range(optim_epochs):
perm = np.arange(actions.size(0))
np.random.shuffle(perm)
perm = torch.LongTensor(perm)
states = states[perm]
actions = actions[perm]
values = values[perm]
targets = targets[perm]
advantages = advantages[perm]
perm_exp = np.arange(expert_actions.size(0))
np.random.shuffle(perm_exp)
perm_exp = torch.LongTensor(perm_exp)
expert_states = expert_states[perm_exp]
expert_actions = expert_actions[perm_exp]
cur_id = 0
cur_id_exp = 0
for _ in range(optim_iters):
cur_batch_size = min(optim_batch_size, actions.size(0) - cur_id)
cur_batch_size_exp = min(optim_batch_size_exp, expert_actions.size(0) - cur_id_exp)
state_var = Variable(states[cur_id:cur_id+cur_batch_size])
action_var = Variable(actions[cur_id:cur_id+cur_batch_size])
advantages_var = Variable(advantages[cur_id:cur_id+cur_batch_size])
expert_state_var = Variable(expert_states[cur_id_exp:cur_id_exp+cur_batch_size_exp])
expert_action_var = Variable(expert_actions[cur_id_exp:cur_id_exp+cur_batch_size_exp])
# update reward net
opt_reward.zero_grad()
# backprop with expert demonstrations
o = reward_net(torch.cat((expert_state_var, expert_action_var),1))
loss = criterion(o, Variable(torch.zeros(expert_action_var.size(0),1)))
loss.backward()
# backprop with generated demonstrations
o = reward_net(torch.cat((state_var, action_var),1))
loss = criterion(o, Variable(torch.ones(action_var.size(0),1)))
loss.backward()
opt_reward.step()
# compute old and new action probabilities
action_means, action_log_stds, action_stds = policy_net(state_var)
log_prob_cur = normal_log_density(action_var, action_means, action_log_stds, action_stds)
action_means_old, action_log_stds_old, action_stds_old = old_policy_net(state_var)
log_prob_old = normal_log_density(action_var, action_means_old, action_log_stds_old, action_stds_old)
# update value net
opt_value.zero_grad()
value_var = value_net(state_var)
value_loss = (value_var - targets[cur_id:cur_id+cur_batch_size]).pow(2.).mean()
value_loss.backward()
opt_value.step()
# update policy net
opt_policy.zero_grad()
ratio = torch.exp(log_prob_cur - log_prob_old) # pnew / pold
surr1 = ratio * advantages_var[:,0]
surr2 = torch.clamp(ratio, 1.0 - clip_epsilon, 1.0 + clip_epsilon) * advantages_var[:,0]
policy_surr = -torch.min(surr1, surr2).mean()
policy_surr.backward()
torch.nn.utils.clip_grad_norm(policy_net.parameters(), 40)
opt_policy.step()
# set new starting point for batch
cur_id += cur_batch_size
cur_id_exp += cur_batch_size_exp
running_state = ZFilter((num_inputs,), clip=5)
#running_reward = ZFilter((1,), demean=False, clip=10)
episode_lengths = []
optim_epochs = args.optim_epochs
optim_batch_size = args.optim_batch_size
expert = Expert(args.expert_path, num_inputs)
print 'Loading expert trajectories ...'
expert.push()
print 'Expert trajectories loaded.'
for i_episode in count(1):
memory = Memory()
num_steps = 0
reward_batch = 0
true_reward_batch = 0
num_episodes = 0
while num_steps < args.batch_size:
state = env.reset()
#state = running_state(state)
reward_sum = 0
true_reward_sum = 0
for t in range(10000): # Don't infinite loop while learning
if args.use_joint_pol_val:
action = select_action_actor_critic(state)
else:
action = select_action(state)
reward = -math.log(reward_net(torch.cat((Variable(torch.from_numpy(state).unsqueeze(0)), action), 1)).data.numpy()[0,0])
action = action.data[0].numpy()
next_state, true_reward, done, _ = env.step(action)
reward_sum += reward
true_reward_sum += true_reward
#next_state = running_state(next_state)
mask = 1
if done:
mask = 0
memory.push(state, np.array([action]), mask, next_state, reward)
if args.render:
env.render()
if done:
break
state = next_state
num_steps += (t-1)
num_episodes += 1
reward_batch += reward_sum
true_reward_batch += true_reward_sum
reward_batch /= num_episodes
true_reward_batch /= num_episodes
gen_batch = memory.sample()
expert_batch = expert.sample(size=args.num_expert_trajs)
if args.use_joint_pol_val:
for _ in range(10):
update_params_actor_critic(gen_batch, expert_batch, i_episode)
else:
update_params(gen_batch, expert_batch, i_episode, optim_epochs, optim_batch_size)
if i_episode % args.log_interval == 0:
print('Episode {}\tLast reward {}\tAverage reward {}\tLast true reward {}\tAverage true reward {:.2f}'.format(
i_episode, reward_sum, reward_batch, true_reward_sum, true_reward_batch))
if i_episode % args.save_interval == 0:
f_w = open('checkpoints/policy_' + str(args.env_name) + '_ep_' + str(i_episode) + '_batch_' + str(args.batch_size) + '_epochs_' + str(args.optim_epochs) + '_exptraj_' + str(args.num_expert_trajs) + '_reward_' + str(true_reward_batch) + '.pth', 'wb')
checkpoint = {'running_state':running_state}
if args.use_joint_pol_val:
checkpoint['policy'] = ac_net
else:
checkpoint['policy'] = policy_net
torch.save(checkpoint, f_w)
if i_episode == args.num_episodes:
break