-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathutils.py
231 lines (198 loc) · 7.08 KB
/
utils.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
from loguru import logger
import numpy as np
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import os
class Noise:
"""
Based on the OpenAI baselines implementation, available at
https://github.com/openai/baselines/blob/master/baselines/ddpg/noise.py
"""
def __init__(self, mu, sigma=0.1, theta=0.15, dt=0.05, x0=None):
self.theta = theta
self.mu = mu
self.sigma = sigma
self.dt = dt
self.x0 = x0
self.x_prev = np.zeros_like(self.mu)
def get_noise(self):
x = self.x_prev + self.theta * (self.mu - self.x_prev) * self.dt + self.sigma * np.sqrt(self.dt) * np.random.normal(
size=self.mu.shape)
self.x_prev = x
return x
def print_three_models(assigns):
print_list_1 = []
print_list_2 = []
print_list_3 = []
for a_i in range(len(assigns.keys())):
if a_i in assigns.keys():
if a_i < 20:
print_list_1.append(assigns[a_i])
elif a_i >= 20 and a_i < 40:
print_list_2.append(assigns[a_i])
elif a_i >= 40 and a_i < 60:
print_list_3.append(assigns[a_i])
n_comps = len(np.unique(print_list_1+print_list_2+print_list_3))
logger.info("number of cluster: {}", n_comps)
logger.info("DP assignments:")
if len(print_list_1) > 0:
print('model 0: ', print_list_1)
if len(print_list_2) > 0:
print('model 1: ', print_list_2)
if len(print_list_3) > 0:
print('model 2: ', print_list_3)
return n_comps
def plot_reward(subtask_list, rewards, name='Subtask Reward', xlabel='subtask index', y=196, scatter=True):
plt.title(name)
plt.axhline(y=y)
if scatter:
plt.plot(subtask_list, rewards, 'r.', markersize=7)
else:
plt.plot(subtask_list, rewards, 'r', linewidth=2)
plt.ylabel('reward')
plt.xlabel(xlabel)
plt.savefig('./misc/' + name)
# plt.show()
plt.close()
def plot_index(predict_list, label_list, name='Cluster_Results.png'):
plt.figure(figsize=(9,5))
plt.subplot(211)
plt.title(name)
plt.plot(range(len(predict_list)), predict_list, 'r.', markersize=10, label='predict')
plt.ylabel('Component Index')
plt.legend()
plt.subplot(212)
plt.plot(range(len(predict_list)), label_list, 'g.', markersize=10, label='groundtruth')
plt.xlabel('Data Index')
plt.ylabel('Component Index')
plt.legend()
plt.savefig('./misc/' + name)
#plt.show()
plt.close()
def plot_index_results(data, predict_list, label_list):
plt.figure(figsize=(9,5))
plt.subplot(211)
plt.title('Cluster Results with GPytorch')
plt.plot(range(data.shape[0]), predict_list, 'r.', markersize=10, label='predict')
plt.ylabel('Component Index')
plt.legend()
plt.subplot(212)
plt.plot(range(data.shape[0]), label_list, 'g.', markersize=10, label='groundtruth')
plt.xlabel('Data Index')
plt.ylabel('Component Index')
plt.legend()
plt.savefig('./misc/Cluster Results with GPytorch')
plt.show()
def plot_gibbs_time(data, time_record):
plt.figure()
plt.plot(range(data.shape[0]), time_record)
plt.xlabel('Data Index')
plt.ylabel('Time per Datapoint (s)')
plt.title('Cost time with GPytorch')
plt.savefig('./misc/Cost time with GPytorch')
plt.show()
def log_plot(path):
data = np.load(path)
def check_create_dir(path):
if not os.path.isdir(path):
os.makedirs(path)
def dumb_assignment_predict(path, PREFIX=''):
assignment_result = np.load(path, allow_pickle=True)[0]
predict = assignment_result['predict']
label = assignment_result['label']
name = PREFIX + '_assignment_predict.png'
plot_index(predict, label, name)
def dumb_predict_error_plot(path, PREFIX='', ylim=[0, 500]):
data_list = np.load(path, allow_pickle=True)
err_list = []
err_list_std = []
err_train = []
err_train_std = []
for data in data_list:
err_list.append(data["mse_error"])
err_list_std.append(data["mse_error_std"])
err_train.append(data["mse_error_train"])
err_train_std.append(data["mse_error_train_std"])
name = PREFIX + '_predict_error'
plt.figure(figsize=(6,4))
plt.title(name)
plt.plot(range(len(err_list)), err_list, linewidth=1, color='royalblue', label='Test Error')
lower = [a-b for a, b in zip(err_list, err_list_std)]
upper = [a+b for a, b in zip(err_list, err_list_std)]
plt.fill_between(range(len(err_list)), lower, upper, facecolor='royalblue', alpha=0.3)
plt.plot(range(len(err_train)), err_train, linewidth=1, color='deeppink', label='Train Error')
lower = [a-b for a, b in zip(err_train, err_train_std)]
upper = [a+b for a, b in zip(err_train, err_train_std)]
plt.fill_between(range(len(err_train)), lower, upper, facecolor='deeppink', alpha=0.3)
plt.legend()
plt.grid()
plt.ylabel('error')
plt.xlabel('timestep')
plt.ylim(ylim)
plt.savefig('./misc/' + name)
plt.close('all')
def dumb_reward_plot(path, PREFIX='', xlim=[0, 200], ylim=[0, 200], y_line1=1, y_line2=175):
# path = './misc/log/April-14-19-55.npy'
data_list = np.load(path, allow_pickle=True)
name = PREFIX + '_task_reward'
plt.figure(figsize=(6,4))
plt.title(name)
if y_line1:
plt.axhline(y=y_line1, color='k')
acc_r_list = []
for data in data_list:
obs = data["obs"]
rewards = data["rewards"]
acc_r = data["reward_sum"]
acc_r_list.append(acc_r)
num_step = len(rewards)
plt.plot(range(rewards.shape[0]), rewards, linewidth=1, label=str(round(acc_r, 4)))
plt.ylabel('reward')
plt.xlabel('timestep')
plt.xlim(xlim)
plt.ylim([-5, 5])
# plt.legend()
plt.savefig('./misc/' + name)
plt.close('all')
# plt.show()
plt.figure(figsize=(6, 4))
name = PREFIX + '_acc_task_reward'
plt.title(name)
if y_line2:
plt.axhline(y=y_line2)
plt.plot(range(len(acc_r_list)), acc_r_list, 'C2', linewidth=2)
plt.grid()
'''
color_bar = ['green', 'yellow', 'red', 'blue']
p_1 = [0, 1, 2, 12, 13, 14, 24, 25, 26, 36, 37, 38]
p_2 = [3, 4, 5, 15, 16, 17, 27, 28, 29, 39, 40, 41]
p_3 = [6, 7, 8, 18, 19, 20, 30, 31, 32, 42, 43, 44]
p_4 = [9, 10, 11, 21, 22, 23, 33, 34, 35, 45, 46, 47]
for p_i in range(1, len(data_list)):
if p_i in p_1:
c_i = color_bar[0]
elif p_i in p_2:
c_i = color_bar[1]
elif p_i in p_3:
c_i = color_bar[2]
else:
c_i = color_bar[3]
plt.fill_between([p_i-1, p_i], ylim[0], ylim[1], facecolor=c_i, alpha=0.2)
'''
plt.ylim(ylim)
plt.ylabel('acc_reward')
plt.xlabel('episode')
# plt.legend()
plt.savefig('./misc/' + name)
plt.close('all')
# plt.show()
def save_test_succ_rate(succ_rate_list, PREFIX=''):
plt.figure(figsize=(6, 4))
name = PREFIX + '_test_acc_reward'
plt.title(name)
plt.axhline(y=1.0)
plt.plot(range(len(succ_rate_list)), succ_rate_list, 'C2', linewidth=2)
plt.grid()
plt.savefig('./misc/' + name)
plt.close('all')