-
Notifications
You must be signed in to change notification settings - Fork 0
/
07_dqn_distrib.py
executable file
·235 lines (186 loc) · 8.2 KB
/
07_dqn_distrib.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
#!/usr/bin/env python3
import gym
import ptan
import numpy as np
import argparse
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from tensorboardX import SummaryWriter
from lib import common
SAVE_STATES_IMG = False
SAVE_TRANSITIONS_IMG = False
if SAVE_STATES_IMG or SAVE_TRANSITIONS_IMG:
import matplotlib as mpl
mpl.use("Agg")
import matplotlib.pylab as plt
Vmax = 10
Vmin = -10
N_ATOMS = 51
DELTA_Z = (Vmax - Vmin) / (N_ATOMS - 1)
STATES_TO_EVALUATE = 1000
EVAL_EVERY_FRAME = 100
class DistributionalDQN(nn.Module):
def __init__(self, input_shape, n_actions):
super(DistributionalDQN, self).__init__()
self.conv = nn.Sequential(
nn.Conv2d(input_shape[0], 32, kernel_size=8, stride=4),
nn.ReLU(),
nn.Conv2d(32, 64, kernel_size=4, stride=2),
nn.ReLU(),
nn.Conv2d(64, 64, kernel_size=3, stride=1),
nn.ReLU()
)
conv_out_size = self._get_conv_out(input_shape)
self.fc = nn.Sequential(
nn.Linear(conv_out_size, 512),
nn.ReLU(),
nn.Linear(512, n_actions * N_ATOMS)
)
self.register_buffer("supports", torch.arange(Vmin, Vmax+DELTA_Z, DELTA_Z))
self.softmax = nn.Softmax(dim=1)
def _get_conv_out(self, shape):
o = self.conv(torch.zeros(1, *shape))
return int(np.prod(o.size()))
def forward(self, x):
batch_size = x.size()[0]
fx = x.float() / 256
conv_out = self.conv(fx).view(batch_size, -1)
fc_out = self.fc(conv_out)
return fc_out.view(batch_size, -1, N_ATOMS)
def both(self, x):
cat_out = self(x)
probs = self.apply_softmax(cat_out)
weights = probs * self.supports
res = weights.sum(dim=2)
return cat_out, res
def qvals(self, x):
return self.both(x)[1]
def apply_softmax(self, t):
return self.softmax(t.view(-1, N_ATOMS)).view(t.size())
def calc_values_of_states(states, net, device="cpu"):
mean_vals = []
for batch in np.array_split(states, 64):
states_v = torch.tensor(batch).to(device)
action_values_v = net.qvals(states_v)
best_action_values_v = action_values_v.max(1)[0]
mean_vals.append(best_action_values_v.mean().item())
return np.mean(mean_vals)
def save_state_images(frame_idx, states, net, device="cpu", max_states=200):
ofs = 0
p = np.arange(Vmin, Vmax + DELTA_Z, DELTA_Z)
for batch in np.array_split(states, 64):
states_v = torch.tensor(batch).to(device)
action_prob = net.apply_softmax(net(states_v)).data.cpu().numpy()
batch_size, num_actions, _ = action_prob.shape
for batch_idx in range(batch_size):
plt.clf()
for action_idx in range(num_actions):
plt.subplot(num_actions, 1, action_idx+1)
plt.bar(p, action_prob[batch_idx, action_idx], width=0.5)
plt.savefig("states/%05d_%08d.png" % (ofs + batch_idx, frame_idx))
ofs += batch_size
if ofs >= max_states:
break
def save_transition_images(batch_size, predicted, projected, next_distr, dones, rewards, save_prefix):
for batch_idx in range(batch_size):
is_done = dones[batch_idx]
reward = rewards[batch_idx]
plt.clf()
p = np.arange(Vmin, Vmax + DELTA_Z, DELTA_Z)
plt.subplot(3, 1, 1)
plt.bar(p, predicted[batch_idx], width=0.5)
plt.title("Predicted")
plt.subplot(3, 1, 2)
plt.bar(p, projected[batch_idx], width=0.5)
plt.title("Projected")
plt.subplot(3, 1, 3)
plt.bar(p, next_distr[batch_idx], width=0.5)
plt.title("Next state")
suffix = ""
if reward != 0.0:
suffix = suffix + "_%.0f" % reward
if is_done:
suffix = suffix + "_done"
plt.savefig("%s_%02d%s.png" % (save_prefix, batch_idx, suffix))
def calc_loss(batch, net, tgt_net, gamma, device="cpu", save_prefix=None):
states, actions, rewards, dones, next_states = common.unpack_batch(batch)
batch_size = len(batch)
states_v = torch.tensor(states).to(device)
actions_v = torch.tensor(actions).to(device)
next_states_v = torch.tensor(next_states).to(device)
# next state distribution
next_distr_v, next_qvals_v = tgt_net.both(next_states_v)
next_actions = next_qvals_v.max(1)[1].data.cpu().numpy()
next_distr = tgt_net.apply_softmax(next_distr_v).data.cpu().numpy()
next_best_distr = next_distr[range(batch_size), next_actions]
dones = dones.astype(np.bool)
# project our distribution using Bellman update
proj_distr = common.distr_projection(next_best_distr, rewards, dones, Vmin, Vmax, N_ATOMS, gamma)
# calculate net output
distr_v = net(states_v)
state_action_values = distr_v[range(batch_size), actions_v.data]
state_log_sm_v = F.log_softmax(state_action_values, dim=1)
proj_distr_v = torch.tensor(proj_distr).to(device)
if save_prefix is not None:
pred = F.softmax(state_action_values, dim=1).data.cpu().numpy()
save_transition_images(batch_size, pred, proj_distr, next_best_distr, dones, rewards, save_prefix)
loss_v = -state_log_sm_v * proj_distr_v
return loss_v.sum(dim=1).mean()
if __name__ == "__main__":
params = common.HYPERPARAMS['pong']
# params['epsilon_frames'] *= 2
parser = argparse.ArgumentParser()
parser.add_argument("--cuda", default=False, action="store_true", help="Enable cuda")
args = parser.parse_args()
device = torch.device("cuda" if args.cuda else "cpu")
env = gym.make(params['env_name'])
env = ptan.common.wrappers.wrap_dqn(env)
writer = SummaryWriter(comment="-" + params['run_name'] + "-distrib")
net = DistributionalDQN(env.observation_space.shape, env.action_space.n).to(device)
tgt_net = ptan.agent.TargetNet(net)
selector = ptan.actions.EpsilonGreedyActionSelector(epsilon=params['epsilon_start'])
epsilon_tracker = common.EpsilonTracker(selector, params)
agent = ptan.agent.DQNAgent(lambda x: net.qvals(x), selector, device=device)
exp_source = ptan.experience.ExperienceSourceFirstLast(env, agent, gamma=params['gamma'], steps_count=1)
buffer = ptan.experience.ExperienceReplayBuffer(exp_source, buffer_size=params['replay_size'])
optimizer = optim.Adam(net.parameters(), lr=params['learning_rate'])
frame_idx = 0
eval_states = None
prev_save = 0
save_prefix = None
with common.RewardTracker(writer, params['stop_reward']) as reward_tracker:
while True:
frame_idx += 1
buffer.populate(1)
epsilon_tracker.frame(frame_idx)
new_rewards = exp_source.pop_total_rewards()
if new_rewards:
if reward_tracker.reward(new_rewards[0], frame_idx, selector.epsilon):
break
if len(buffer) < params['replay_initial']:
continue
if eval_states is None:
eval_states = buffer.sample(STATES_TO_EVALUATE)
eval_states = [np.array(transition.state, copy=False) for transition in eval_states]
eval_states = np.array(eval_states, copy=False)
optimizer.zero_grad()
batch = buffer.sample(params['batch_size'])
save_prefix = None
if SAVE_TRANSITIONS_IMG:
interesting = any(map(lambda s: s.last_state is None or s.reward != 0.0, batch))
if interesting and frame_idx // 30000 > prev_save:
save_prefix = "images/img_%08d" % frame_idx
prev_save = frame_idx // 30000
loss_v = calc_loss(batch, net, tgt_net.target_model, gamma=params['gamma'],
device=device, save_prefix=save_prefix)
loss_v.backward()
optimizer.step()
if frame_idx % params['target_net_sync'] == 0:
tgt_net.sync()
if frame_idx % EVAL_EVERY_FRAME == 0:
mean_val = calc_values_of_states(eval_states, net, device=device)
writer.add_scalar("values_mean", mean_val, frame_idx)
if SAVE_STATES_IMG and frame_idx % 10000 == 0:
save_state_images(frame_idx, eval_states, net, device=device)