-
Notifications
You must be signed in to change notification settings - Fork 188
/
plot_cer.py
executable file
·116 lines (93 loc) · 4.53 KB
/
plot_cer.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
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import sys
import matplotlib
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
plotfile = sys.argv[1]
modelname = sys.argv[2]
ytsvfile = sys.argv[3] # "iteration.tsv"
ctsvfile = sys.argv[4] # "checkpoint.tsv"
etsvfile = sys.argv[5] # "eval.tsv" - Not used as no training iterations number
stsvfile = sys.argv[6] # "sub.tsv"
ltsvfile = sys.argv[7] # "lstmeval.tsv"
maxticks=10
ydf = pd.read_csv(ytsvfile,sep='\t', encoding='utf-8')
cdf = pd.read_csv(ctsvfile,sep='\t', encoding='utf-8')
sdf = pd.read_csv(stsvfile,sep='\t', encoding='utf-8')
ldf = pd.read_csv(ltsvfile,sep='\t', encoding='utf-8')
ydf = ydf.sort_values('TrainingIteration')
cdf = cdf.sort_values('TrainingIteration')
sdf = sdf.sort_values('TrainingIteration')
ldf = ldf.sort_values('TrainingIteration')
y = ydf['IterationCER']
x = ydf['LearningIteration']
t = ydf['TrainingIteration']
c = cdf['CheckpointCER']
cx = cdf['LearningIteration']
ct = cdf['TrainingIteration']
s = sdf['SubtrainerCER']
sx = sdf['LearningIteration']
st = sdf['TrainingIteration']
l = ldf['EvalCER']
lx = ldf['LearningIteration']
lt = ldf['TrainingIteration']
def annot_min(boxcolor, xpos, ypos, x, y, z):
tmin = z.iloc[np.argmin(y)]
xmin = x.iloc[np.argmin(y)]
ymin = y.min()
boxtext= " {:.3f}% at {:,} / {:,} " .format(ymin,xmin,tmin)
ax1.annotate(boxtext, xy=(tmin, ymin), xytext=(xpos,ypos), textcoords='offset points', color='black', fontsize=9,
arrowprops=dict(shrinkA=1, shrinkB=1, fc=boxcolor,alpha=0.7, ec='white', connectionstyle="arc3"),
bbox=dict(boxstyle='round,pad=0.2', fc=boxcolor, alpha=0.3))
PlotTitle="Tesseract LSTM Training : " + modelname
fig = plt.figure(figsize=(11,8.5)) #size is in inches
ax1 = fig.add_subplot()
ax1.yaxis.set_major_formatter(matplotlib.ticker.ScalarFormatter())
ax1.yaxis.set_major_formatter(matplotlib.ticker.FormatStrFormatter("%.1f"))
ax1.set_ylabel('Error Rate %')
ax1.set_xlabel('Training Iterations')
ax1.set_xticks(t)
ax1.tick_params(axis='x', labelsize='small')
ax1.locator_params(axis='x', nbins=maxticks) # limit ticks on x-axis
ax1.xaxis.set_major_formatter(matplotlib.ticker.ScalarFormatter())
ax1.xaxis.set_major_formatter(matplotlib.ticker.StrMethodFormatter('{x:,.0f}'))
ax1.scatter(t, y, c='teal', alpha=0.7, s=0.5, label='BCER at #iterations/100 - lstmtraining - list.train')
ax1.plot(t, y, 'teal', alpha=0.3, linewidth=0.5, label='Training BCER')
ax1.grid(True)
if not c.dropna().empty: # not NaN or empty
ax1.scatter(ct, c, c='teal', marker='x', s=35,
label='BCER at checkpoints - lstmtraining - list.train', alpha=0.5)
annot_min('teal',0,-50,cx,c,ct)
if not l.dropna().empty: # not NaN or empty
ax1.plot(lt, l, 'magenta', linewidth=0.5, label='Validation BCER')
ax1.scatter(lt, l, c='magenta', s=10, label='BCER at checkpoints - lstmeval - list.eval', alpha=0.5)
annot_min('magenta',0,-50,lx,l,lt)
if not s.dropna().empty: # not NaN or empty
ax1.plot(st, s, 'orange', linewidth=0.5, label='SubTrainer BCER')
ax1.scatter(st, s, c='orange', s=0.5,
label='BCER for UpdateSubtrainer every 100 iterations', alpha=0.5)
annot_min('orange',-100,-100,sx,s,st)
tmax = t[np.argmax(x)]
ymax = y[np.argmax(x)]
xmax = x.max()
boxtext= " {:.3f}% at \n {:,} \n {:,} " .format(ymax,xmax,tmax)
ax1.annotate(boxtext, xy=(tmax, ymax), xytext=(20,-10), textcoords='offset points', color='black',
bbox=dict(boxstyle='round,pad=0.2', fc='teal', alpha=0.3))
plt.title('character error rate over training iterations - from lstmtraining and lstmeval',fontsize=10)
plt.suptitle(PlotTitle, y=0.95, fontsize = 14, fontweight = 'bold')
plt.legend(loc='upper right')
ax1.set_ylim([-0.5,100])
# Secondary x axis on top to display Learning Iterations
ax2 = ax1.twiny() # ax1 and ax2 share y-axis
ax2.set_xlabel("Learning Iterations")
ax2.set_xlim(ax1.get_xlim()) # ensure the independent x-axes now span the same range
ax2.set_xticks(t) # copy over the locations of the x-ticks from Training Iterations
ax2.tick_params(axis='x', labelsize='small')
ax2.set_xticklabels(matplotlib.ticker.StrMethodFormatter('{x:,.0f}').format_ticks(x)) # But give value of Learning Iterations
ax2.locator_params(axis='x', nbins=maxticks) # limit ticks to same as x-axis
ax2.xaxis.set_ticks_position('bottom') # set the position of ticks of the second x-axis to bottom
ax2.xaxis.set_label_position('bottom') # set the position of labels of the second x-axis to bottom
ax2.spines['bottom'].set_position(('outward', 36)) # positions the second x-axis below the first x-axis
plt.savefig(plotfile)