-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathval.py
158 lines (125 loc) · 6.88 KB
/
val.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
import torch
import torch.nn as nn
import torchvision.transforms as transforms
from license_plate_dataset import LicensePlateDataset, LicensePlateVocab
from license_plate_model import LicensePlateModel
from torch.utils.data import DataLoader
import matplotlib.pyplot as plt
import torch.optim as optim
import matplotlib.font_manager as fm
from tqdm import tqdm
import os
import argparse
img_height = 224
img_width = 224
# 最大序列长度
max_length = 16
num_epochs = 100
# 设置中文字体
font_path = "C:/Windows/Fonts/simhei.ttf" # 可以选择其他中文字体
prop = fm.FontProperties(fname=font_path)
# 词汇表
vocab_list = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'J', 'K', 'L', 'M', 'N', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', '云', '京', '冀', '吉', '学', '宁', '川', '挂', '新', '晋', '桂', '沪', '津', '浙', '渝', '湘', '琼', '甘', '皖', '粤', '苏', '蒙', '藏', '警', '豫', '贵', '赣', '辽', '鄂', '闽', '陕', '青', '鲁', '黑']
vocab = LicensePlateVocab(vocab_list)
def generate_license_plate(model, image, vocab, device, max_length=16):
with torch.no_grad():
memory = model.encode_image(image)
expression = ''
generated_tokens = vocab.text_to_sequence(expression, max_length=max_length, pad_to_max_length=False, add_eos=False, add_bos=True)
while True:
tgt_input = torch.tensor(generated_tokens).unsqueeze(0).to(device)
output = model.decode_text(memory, tgt_input)
next_token = output.argmax(dim=-1)[:, -1].item()
if next_token == vocab.eos_idx or len(expression) == max_length:
break
expression += vocab.sequence_to_text([next_token])
generated_tokens = vocab.text_to_sequence(expression, max_length=max_length, pad_to_max_length=False, add_eos=False, add_bos=True)
return expression
def generate_license_plate_once(model, image, vocab, device, max_length=16):
with torch.no_grad():
memory = model.encode_image(image)
expression = ''
generated_tokens = vocab.text_to_sequence(expression, max_length=max_length, pad_to_max_length=True, add_eos=False, add_bos=True)
tgt_input = torch.tensor(generated_tokens).unsqueeze(0).to(device)
output = model.decode_text(memory, tgt_input)
next_token = output.argmax(dim=-1).cpu().numpy()[0]
# 裁剪掉EOS之后的数据
if vocab.eos_idx in next_token:
eos_index = list(next_token).index(vocab.eos_idx)
next_token = next_token[:eos_index]
expression = vocab.sequence_to_text(next_token)
return expression
# 验证函数
def validate(model, val_loader, criterion, device):
model.eval()
val_loss = 0.0
correct = 0
exact_match_correct = 0
total = 0
images_count = 0
batch_idx = 0
val_loader_tqdm = tqdm(val_loader, desc='Validation', unit='batch', leave=False)
with torch.no_grad():
for images, labels in val_loader_tqdm:
images = images.to(device)
labels = labels.to(device)
outputs = model(images, labels[:, :-1])
outputs = outputs.permute(0, 2, 1)
loss = criterion(outputs, labels[:, 1:])
val_loss += loss.item() * images.size(0)
images_count += images.size(0)
_, predicted = torch.max(outputs, 1)
# 创建一个掩码，标识哪些位置不是 vocab.pad_idx
non_pad_mask = (labels[:, 1:] != vocab.pad_idx)
pad_mask = (labels[:, 1:] == vocab.pad_idx)
predicted[pad_mask] = vocab.pad_idx
# 使用掩码来计算正确的预测
correct += (predicted[non_pad_mask] == labels[:, 1:][non_pad_mask]).sum().item()
total += non_pad_mask.sum().item()
# 计算每个序列的完全匹配情况
exact_match_correct += ((predicted == labels[:, 1:]) | ~non_pad_mask).all(dim=1).sum().item()
val_loader_tqdm.set_postfix(loss=f"{val_loss/images_count:.5f}")
# Plotting the results
fig, axes = plt.subplots(10, 10, figsize=(15, 15))
axes = axes.flatten()
for i, (image, label, pred) in enumerate(zip(images, labels, predicted)):
image_np = image.cpu().numpy().transpose(1, 2, 0)
real_plate = vocab.sequence_to_text(label.cpu().numpy()[1:]) # Skip the start token
predicted_plate = vocab.sequence_to_text(pred.cpu().numpy())
axes[i].imshow(image_np)
if real_plate != predicted_plate:
axes[i].set_title(f'标注: {real_plate}\n预测: {predicted_plate}', fontproperties=prop, color='red')
else:
axes[i].set_title(f'标注: {real_plate}\n预测: {predicted_plate}', fontproperties=prop)
axes[i].axis('off')
plt.tight_layout()
os.makedirs('output_images', exist_ok=True)
plt.savefig(f'output_images/generated_plates_{batch_idx:05d}.png')
batch_idx += 1
plt.close()
val_loss = val_loss / images_count
accuracy = correct / total
exact_match_accuracy = exact_match_correct / images_count
return val_loss, accuracy, exact_match_accuracy
if __name__ == '__main__':
parser = argparse.ArgumentParser(description="Validate a license plate recognition model")
parser.add_argument('--val_folder', type=str, default=r'./datasets/test', help='Path to the validation dataset folder')
parser.add_argument('--checkpoint_path', type=str, default=r'./checkpoints/last_model.pth', help='Path to the model checkpoint file')
args = parser.parse_args()
# 设置数据变换
transform = transforms.Compose([
transforms.Resize((img_height, img_width)),
transforms.ToTensor()
])
# 创建数据集和数据加载器
val_dataset = LicensePlateDataset(args.val_folder, vocab, max_length, transform)
val_loader = DataLoader(val_dataset, batch_size=100, shuffle=True, num_workers=4)
data_iter = iter(val_loader)
images, labels = next(data_iter)
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
# model = LicensePlateModel(pad_idx=vocab.pad_idx, vocab_size=len(vocab.vocab_list), max_length=max_length)
model = LicensePlateModel(pad_idx=vocab.pad_idx, d_model=64, nhead_encoder=4, nhead_decoder=4, num_encoder_layers=2, num_decoder_layers=2, dim_feedforward=512, vocab_size=len(vocab.vocab_list), max_length=max_length)
model.load_state_dict(torch.load(args.checkpoint_path)['model_state_dict'])
model.to(device)
val_loss, accuracy, exact_match_accuracy = validate(model, val_loader, nn.CrossEntropyLoss(ignore_index=vocab.pad_idx), device)
print(f"Validation Loss: {val_loss:.5f}, Accuracy: {accuracy:.5f}, Exact Match Accuracy: {exact_match_accuracy:.5f}")