-
Notifications
You must be signed in to change notification settings - Fork 30
/
app.py
139 lines (118 loc) · 5.09 KB
/
app.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
# -*- coding: utf-8 -*-
import sys
sys.path.insert(0, 'src')
# import the necessary packages
import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='3'
import pickle
import random
import cv2 as cv
import keras.backend as K
import numpy as np
from keras.preprocessing import sequence
#从config文件中引入一些参数 包括token最大长度 测试文件夹长度 最优的模型参数
from config import max_token_length, test_a_image_folder, best_model
from forward import build_model
from generated import test_gen
#使用训练好的模型对图片进行测试
def beam_search_predictions(model, image_name, word2idx, idx2word, encoding_test, beam_index=3):
start = [word2idx["<start>"]]
start_word = [[start, 0.0]]
while len(start_word[0][0]) < max_token_length:
temp = []
for s in start_word:
#对序列进行填充的预处理，在其后添0，使其序列统一大小为max_token_length
par_caps = sequence.pad_sequences([s[0]], maxlen=max_token_length, padding='post')
#每次取一个图片进行测试
e = encoding_test[image_name]
#使用模型对该图片进行测试
preds = model.predict([np.array([e]), np.array(par_caps)])
#从预测的结果中取前beam_index个
word_preds = np.argsort(preds[0])[-beam_index:]
# Getting the top <beam_index>(n) predictions and creating a
# new list so as to put them via the model again
# 创建一个新的list结构 将预测出的词和词的概率以组对的形式存储
for w in word_preds:
next_cap, prob = s[0][:], s[1]
next_cap.append(w)
prob += preds[0][w]
temp.append([next_cap, prob])
#将处理好的预测值赋值回start word
start_word = temp
# Sorting according to the probabilities
# 根据概率排序
start_word = sorted(start_word, reverse=False, key=lambda l: l[1])
# Getting the top words
#获得最有可能正确的词
start_word = start_word[-beam_index:]
start_word = start_word[-1][0]
#根据id取出单词
intermediate_caption = [idx2word[i] for i in start_word]
final_caption = []
#组合成句
for i in intermediate_caption:
if i != '<end>':
final_caption.append(i)
else:
break
final_caption = ''.join(final_caption[1:])
return final_caption
if __name__ == '__main__':
#图片的channel为3
channel = 3
#设置模型权重的地址
model_weights_path = os.path.join('models', best_model)
print('模型加载中...')
#创建模型
model = build_model()
#加载模型权重
model.load_weights(model_weights_path)
print('模型加载完毕...')
#print(model.summary())
#加载语料库
vocab = pickle.load(open('data/vocab_train.p', 'rb'))
#将word转化为数字 方便输入网络 进行预测
idx2word = sorted(vocab)
word2idx = dict(zip(idx2word, range(len(vocab))))
print('语料库加载完毕...')
test_gen()
#加载测试图片
encoding_test = pickle.load(open('data/encoded_test_a_images.p', 'rb'))
#随机取测试图片
names = [f for f in encoding_test.keys()]
samples = names
#samples = random.sample(names, 1)
sentences = []
for i in range(len(samples)):
image_name = samples[i]
image_input = np.zeros((1, 2048))
image_input[0] = encoding_test[image_name]
#获取图片的名称
filename = os.path.join(test_a_image_folder, image_name)
# print('Start processing image: {}'.format(filename))
#设置不同的预测参数，并放到beam_search_predictions中进行预测
print('描述的图片为:',image_name)
candidate1=beam_search_predictions(model, image_name, word2idx, idx2word, encoding_test, beam_index=1)
print('Beam Search, k=1:',candidate1)
sentences.append(candidate1)
candidate2=beam_search_predictions(model, image_name, word2idx, idx2word, encoding_test, beam_index=3)
print('Beam Search, k=3:',candidate2)
sentences.append(candidate2)
candidate3=beam_search_predictions(model, image_name, word2idx, idx2word, encoding_test, beam_index=5)
print('Beam Search, k=5:',candidate3)
sentences.append(candidate3)
candidate4=beam_search_predictions(model, image_name, word2idx, idx2word, encoding_test, beam_index=7)
print('Beam Search, k=7:',candidate4)
sentences.append(candidate4)
#读取图片
img = cv.imread(filename)
#resise到固定大小 256*256
#img = cv.resize(img, (256, 256), cv.INTER_CUBIC)
if not os.path.exists('images'):
os.makedirs('images')
#将修改后的图片重新写回
cv.imwrite('images/{}_bs_image.jpg'.format(i), img)
#将预测产生的描述信息输出到demo.txt文件中
with open('demo.txt', 'w') as file:
file.write('\n'.join(sentences))
K.clear_session()