/
validate_network.py
105 lines (91 loc) · 3.25 KB
/
validate_network.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
import sys
import json
import torchvision.transforms as transforms
import torch
from torch.utils.data import DataLoader
from tqdm import tqdm
import editdistance
# ~ from ocr_dataset import OCRDataset
from orfac_dataset import ORFACDataset
# ~ from ocr_dataset import collate_batch
from orfac_dataset import orfac_collate_batch
from network import OCROnly
from converter import Converter
def validate_loaded_net(socr, data_path):
device = 'cuda:0'
cs = json.load(open('charmap.json', 'rt'))
cs['/PAD/'] = len(cs)+1
trans = transforms.Compose([transforms.Grayscale(), transforms.ToTensor()])
dataset = ORFACDataset(folder=data_path, charset=None, transform=trans)
converter = Converter(cs)
test_dataloader = DataLoader(dataset, batch_size=1, shuffle=False, collate_fn=orfac_collate_batch)
socr = socr.to(device)
with torch.no_grad():
socr.eval()
d_sum = 0
c_sum = 0
for tns, base_width, lbl, _, cf, pf in test_dataloader:
tns = tns.to(device)
out = socr(tns)
out = out.transpose(0,1)
am = torch.argmax(out[:, :, :], 2)
res = converter.decode(am, base_width)
for i in range(len(lbl)):
ed = editdistance.eval(res[i], lbl[i])
ll = len(lbl[i])
d_sum += ed
c_sum += ll
if c_sum==0:
print()
print()
print('Error', data_path)
print()
print()
print()
quit()
return 100*d_sum/c_sum
def validate_network(model_path, data_path, display=False):
device = 'cuda:0'
batch = 32
cs = json.load(open('charmap.json', 'rt'))
cs['/PAD/'] = len(cs)+1
trans = transforms.Compose([transforms.Grayscale(), transforms.ToTensor()])
dataset = ORFACDataset(folder=data_path, charset=None, transform=trans)
if len(dataset)==0:
raise Exception()
network = OCROnly(nb_classes=(len(cs)+1), feature_dim=128).to(device)
try:
# ~ network.load_state_dict(torch.load(model_path, map_location=device))
network.load(model_path)
except:
print('Cannot load network')
quit(1)
converter = Converter(cs)
test_dataloader = DataLoader(dataset, batch_size=1, shuffle=False, collate_fn=orfac_collate_batch)
with torch.no_grad():
network.eval()
d_sum = 0
c_sum = 0
for tns, base_width, lbl, _, cf, pf in test_dataloader:
tns = tns.to(device)
out = network(tns)
out = out.transpose(0,1)
am = torch.argmax(out[:, :, :], 2)
res = converter.decode(am, base_width)
for i in range(len(lbl)):
ed = editdistance.eval(res[i], lbl[i])
ll = len(lbl[i])
d_sum += ed
c_sum += ll
if display:
print()
print(lbl[i])
print(res[i])
print('%d/%d=%.1f%%' % (ed, ll, 100*ed/ll))
if display:
print('CER: %.2f' % (100*d_sum/c_sum))
return 100*d_sum/c_sum
if __name__=='__main__':
model_path = sys.argv[1]
font_group = sys.argv[2]
validate_network(model_path, 'data/valid/single/%s' % font_group, display=True)