-
Notifications
You must be signed in to change notification settings - Fork 0
/
draw_loss.py
96 lines (81 loc) · 3.6 KB
/
draw_loss.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
import os
import pdb
import sys
import re
import copy
import numpy as np
import matplotlib.pyplot as plt
def save(path, ext='png', close=True, verbose=True):
# Extract the directory and filename from the given path
directory = os.path.split(path)[0]
filename = "%s.%s" % (os.path.split(path)[1], ext)
if directory == '':
directory = '.'
# If the directory does not exist, create it
if not os.path.exists(directory):
os.makedirs(directory)
#The final path to save to
savepath = os.path.join(directory, filename)
if verbose:
print("Saving figure to '%s'..." % savepath),
# Actually save the figure
plt.savefig(savepath)
# Close it
if close:
plt.close()
if verbose:
print("Done")
if __name__ == '__main__':
ctc_fname = sys.argv[1]
em_fname = sys.argv[2]
re_dev_str = 'dev set.*, losses = \[.*\]'
re_test_str = 'test set.*, losses = \[.*\]'
re_train_str = 'Training loss = \[.*\]'
for phase, fname, c in zip(['ctc', 'em'], [ctc_fname, em_fname], [['r', 'g', 'b'], ['c', 'm', 'k']]):
losses = []
loss = []
with open(fname) as f:
for line in f.readlines():
for match in re.findall(re_dev_str, line):
if len(loss) > 0:
iteration = int(re.findall('Epoch [0-9]*', line)[0].split(' ')[-1])-1
loss[-1] = np.array(loss[-1]).mean(axis=0)
loss.append(iteration)
losses.append(loss)
wer = float(re.findall('WER = [0-9\.]*', match)[0].split(' ')[-1])*100
match = [wer] + map(float, re.findall('[0-9\.]+', match.split('= ')[-1]))
loss = [match, [], []]
for match in re.findall(re_test_str, line):
wer = float(re.findall('WER = [0-9\.]*', match)[0].split(' ')[-1])*100
match = [wer] + map(float, re.findall('[0-9\.]+', match.split('= ')[-1]))
loss[1] = match
for match in re.findall(re_train_str, line):
match = map(float, re.findall('[0-9\.]+', match.split('= ')[-1]))
# pdb.set_trace()
loss[2].append(match)
iters = [l[-1] for l in losses]
ctc_dev = [l[0][-3] for l in losses]
ctc_test = [l[1][-3] for l in losses]
ctc_train = [l[2][-3] for l in losses]
best_dev = [l[0][-2] for l in losses]
best_test = [l[1][-2] for l in losses]
best_train = [l[2][-2] for l in losses]
greed_dev = [l[0][-1] for l in losses]
greed_test = [l[1][-1] for l in losses]
greed_train = [l[2][-1] for l in losses]
wer_dev = [l[0][0] for l in losses]
wer_test = [l[1][0] for l in losses]
# pdb.set_trace()
# plt.plot(iters, ctc_train, color=c[0], label='%s: ctc_train'%phase)
# plt.plot(iters, best_train, color=c[1], label='%s: best_train'%phase)
# plt.plot(iters, greed_train, color=c[2], label='%s: greed_train'%phase)
plt.plot(iters, ctc_dev, color=c[0], label='%s: ctc_dev'%phase)
plt.plot(iters, best_dev, color=c[1], label='%s: best_dev'%phase)
plt.plot(iters, greed_dev, color=c[2], label='%s: greed_dev'%phase)
plt.plot(iters, wer_dev, color=c[2], label='%s: WER_dev'%phase)
plt.ylim([0, 70])
plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
plt.show()
# handles, labels = plt.get_legend_handles_labels()
# plt.legend(handles=[ct_line, bt_line, gt_line])
save('dev', ext='png', close=True)