-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathfill_between_demo.py
163 lines (139 loc) · 5.94 KB
/
fill_between_demo.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
import numpy as np
import matplotlib.pyplot as plt
from matplotlib import rcParams
from matplotlib.ticker import FuncFormatter
# def mean_list(valid_ids, valid_values, stride=2, iters=1000):
# length = len(valid_ids)
# nums = length // stride
# used_ids = []
# mean_values = []
# std_vallues = []
# valid_values = np.asarray(valid_values)
# for k in range(nums):
# used_ids.append(int(k*stride*iters))
# temp_list = valid_values[k*stride:(k+1)*stride]
# mean = np.mean(temp_list)
# std = np.std(temp_list)
# mean_values.append(mean)
# std_vallues.append(std)
# return used_ids, mean_values, std_vallues
def mean_list(loss1, loss2, loss3):
with open(loss1, 'r', encoding='utf-8') as l1f:
loss1_list = []
for line in l1f.readlines():
line = line.strip()
loss1_list.append(float(line))
with open(loss2, 'r', encoding='utf-8') as l2f:
loss2_list = []
for line in l2f.readlines():
line = line.strip()
loss2_list.append(float(line))
with open(loss3, 'r', encoding='utf-8') as l3f:
loss3_list = []
for line in l3f.readlines():
line = line.strip()
loss3_list.append(float(line))
# totoal num=100, n=1
n = 2
new_loss1_list = []
for i in range(0, len(loss1_list), n):
mean = np.mean(loss1_list[i: i + n])
new_loss1_list.append(mean)
new_loss2_list = []
for i in range(0, len(loss2_list), n):
mean = np.mean(loss2_list[i: i + n])
new_loss2_list.append(mean)
new_loss3_list = []
for i in range(0, len(loss3_list), n):
mean = np.mean(loss3_list[i: i + n])
new_loss3_list.append(mean)
mean_values = []
std_vallues = []
for i, loss_tuple in enumerate(zip(new_loss1_list, new_loss2_list, new_loss3_list)):
loss_list = list(loss_tuple)
mean = np.mean(loss_list)
std = np.std(loss_list)
mean_values.append(mean)
std_vallues.append(std)
# print(len(mean_values))
# print(len(std_vallues))
return mean_values, std_vallues
# 设置全局格式，包括字体风格和大小等等
# 这里主要用来修改文本字体里面的格式
font_size = 50
config = {
"font.family":'serif',
"font.size": font_size,
"mathtext.fontset":'stix',
"font.serif": ['SimSun'],
}
rcParams.update(config)
# 修改x轴的显示方式，科学计数法
def formatnumx(x, pos):
return '%d' % (x/1000)
formatterx = FuncFormatter(formatnumx)
fig, ax1 = plt.subplots(figsize=(16, 16), dpi=100)
# ax2 = ax1.twinx()
debais_method = 'CD_CTDA'
bias_list = ['appearance', 'age', 'gender', 'orientation']
bias = bias_list[0]
if bias == 'appearance':
loss1 = './loss/'+debais_method+'/appearance1.txt'
loss2 = './loss/'+debais_method+'/appearance2.txt'
loss3 = './loss/'+debais_method+'/appearance3.txt'
color = 'C4'
elif bias == 'age':
loss1 = './loss/'+debais_method+'/age1.txt'
loss2 = './loss/'+debais_method+'/age2.txt'
loss3 = './loss/'+debais_method+'/age3.txt'
color = 'C2'
elif bias == 'gender':
loss1 = './loss/'+debais_method+'/gender1.txt'
loss2 = './loss/'+debais_method+'/gender2.txt'
loss3 = './loss/'+debais_method+'/gender3.txt'
color = 'C0'
elif bias == 'orientation':
loss1 = './loss/'+debais_method+'/orientation1.txt'
loss2 = './loss/'+debais_method+'/orientation2.txt'
loss3 = './loss/'+debais_method+'/orientation3.txt'
color = 'C1'
def plot_lines(ap_loss1, ap_loss2, ap_loss3):
ap_mean_values, ap_std_vallues = mean_list(ap_loss1, ap_loss2, ap_loss3)
id = list(range(0, len(ap_mean_values)))
ap_std_down = [ap_mean_values[x]-ap_std_vallues[x] for x in range(len(ap_mean_values))]
ap_std_up = [ap_mean_values[x]+ap_std_vallues[x] for x in range(len(ap_mean_values))]
l1 = ax1.plot(id, ap_mean_values, color=color)
ax1.fill_between(id, ap_std_down, ap_std_up, color=color, alpha=0.3)
# ag_mean_values, ag_std_vallues = mean_list(ag_loss1, ag_loss2, ag_loss3)
# id = list(range(0, len(ag_mean_values)))
# ag_std_down = [ag_mean_values[x] - ag_std_vallues[x] for x in range(len(ag_mean_values))]
# ag_std_up = [ag_mean_values[x] + ag_std_vallues[x] for x in range(len(ag_mean_values))]
# l2 = ax1.plot(id, ag_mean_values, color='C1')
# ax1.fill_between(id, ag_std_down, ag_std_up, color='C1', alpha=0.3)
# plt.legend(handles=[l1, l2], labels=['appearance', 'age'], loc='best')
ap = plot_lines(loss1, loss2, loss3)
# print(len(mean_values))
# valid_ids 为迭代次数（每一千次为一个单位）
# valid_mse 为每1000次的validation值
# used_ids, mean_values, std_vallues = mean_list(valid_ids, valid_mse, stride=2, iters=1000)
# std_down = [mean_values[x]-std_vallues[x] for x in range(len(mean_values))]
# std_up = [mean_values[x]+std_vallues[x] for x in range(len(mean_values))]
# ax2.plot(used_ids, mean_values, color='C1', label='Ours w/o PT')
# ax2.fill_between(used_ids, std_down, std_up, color='C1', alpha=0.3)
ax1.set_xlabel(r'Steps', fontdict={'family': 'Times New Roman', 'size': font_size})
ax1.set_ylabel('Training loss', fontdict={'family': 'Times New Roman', 'size': font_size})
# ax2.set_ylabel('Validation result (VOI)', color='C1', fontdict={'family': 'Times New Roman', 'size': font_size})
#plt.gca().xaxis.set_major_formatter(formatterx)
ax1.tick_params(labelsize=font_size)
ticks = ax1.set_xticks([0, 25, 50, 75, 100])
# print(list(range(0,400,80)))
labels = ax1.set_xticklabels(['0', '100', '200', '300', '400'], rotation=30, fontsize='small')
# labels = ax1.get_xticklabels() + ax1.get_yticklabels()
# [label.set_fontname('Times New Roman') for label in labels]
# ax2.tick_params(labelsize=font_size)
# labels = ax2.get_xticklabels() + ax2.get_yticklabels()
# [label.set_fontname('Times New Roman') for label in labels]
# ax2.set_ylim((1, 6.5))
fname_path = './loss_pic/' + debais_method + '_' + bias + '.pdf'
plt.savefig(fname_path)
# plt.show()