This repository has been archived by the owner on Oct 20, 2024. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 15
/
Copy pathmain.py
172 lines (134 loc) · 5.8 KB
/
main.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
159
160
161
162
163
164
165
166
167
168
169
170
171
172
#!/usr/bin/env python
import sys
import argparse
import tensorflow as tf
import io
from model import OpenNsfwModel, InputType
import flask
from PIL import Image
import numpy as np
import skimage
import skimage.io
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
import os
os.environ["CUDA_VISIBLE_DEVICES"]="-1"
model_weights_path = 'data/open_nsfw-weights.npy'
model = OpenNsfwModel()
VGG_MEAN = [104, 117, 123]
img_width, img_height = 224, 224
app = flask.Flask(__name__)
# 将RGB按照BGR重新组装，然后对每一个RGB对应的值减去一定阈值
def prepare_image(image):
H, W, _ = image.shape
h, w = (img_width, img_height)
h_off = max((H - h) // 2, 0)
w_off = max((W - w) // 2, 0)
image = image[h_off:h_off + h, w_off:w_off + w, :]
image = image[:, :, :: -1]
image = image.astype(np.float32, copy=False)
image = image * 255.0
image = image-np.array(VGG_MEAN, dtype=np.float32)
image = np.expand_dims(image, axis=0)
return image
# 使用TFLite文件检测
def getResultFromFilePathByTFLite(path):
model_path = "./model/nsfw.tflite"
interpreter = tf.lite.Interpreter(model_path=model_path)
interpreter.allocate_tensors()
# Get input and output tensors.
input_details = interpreter.get_input_details()
# print(str(input_details))
output_details = interpreter.get_output_details()
# print(str(output_details))
im = Image.open(path)
# im = Image.open(r"./images/image1.png")
if im.mode != "RGB":
im = im.convert('RGB')
imr = im.resize((256, 256), resample=Image.BILINEAR)
fh_im = io.BytesIO()
imr.save(fh_im, format='JPEG')
fh_im.seek(0)
image = (skimage.img_as_float(skimage.io.imread(fh_im, as_grey=False))
.astype(np.float32))
# 填装数据
final = prepare_image(image)
interpreter.set_tensor(input_details[0]['index'], final)
# 调用模型
interpreter.invoke()
output_data = interpreter.get_tensor(output_details[0]['index'])
# 出来的结果去掉没用的维度
result = np.squeeze(output_data)
print('TFLite-->>result:{},path:{}'.format(result, path))
print(
"==========================================================================================================")
print("")
print("")
def getResultFromFilePathByPyModle(path):
# print("numpy-version:" + np.__version__)
# print("tensorflow-version:" + tf.__version__)
im = Image.open(path)
if im.mode != "RGB":
im = im.convert('RGB')
# print("图片reSize：256*256")
imr = im.resize((256, 256), resample=Image.BILINEAR)
fh_im = io.BytesIO()
imr.save(fh_im, format='JPEG')
fh_im.seek(0)
image = (skimage.img_as_float(skimage.io.imread(fh_im, as_grey=False))
.astype(np.float32))
final = prepare_image(image)
tf.reset_default_graph()
with tf.Session() as sess:
input_type = InputType[InputType.TENSOR.name.upper()]
model.build(weights_path=model_weights_path, input_type=input_type)
sess.run(tf.global_variables_initializer())
predictions = sess.run(model.predictions, feed_dict={model.input: final})
# print("\tSFW score:\t{}\n\tNSFW score:\t{}".format(*predictions[0]))
print(
"==========================================================================================================")
print('Python-->>result:{},path:{}'.format(predictions[0], path))
def getResultListFromDir():
list = os.listdir("/Users/zhaowenwen/Downloads/testImages")
for i in range(0, len(list)):
if (list[i] != ".DS_Store" and list[i] != ".localized"):
getResultFromFilePathByPyModle(os.path.join("/Users/zhaowenwen/Downloads/testImages", list[i]))
getResultFromFilePathByTFLite(os.path.join("/Users/zhaowenwen/Downloads/testImages", list[i]))
# 代码生成tflite文件
def createTfliteFile():
in_path = "./model/frozen_nsfw.pb"
out_path = "./model/nsfw.tflite"
# 模型输入节点
input_tensor_name = ["input"]
input_tensor_shape = {"input":[1, 224,224,3]}
# 模型输出节点
classes_tensor_name = ["predictions"]
converter = tf.lite.TFLiteConverter.from_frozen_graph(in_path,
input_tensor_name, classes_tensor_name,
input_shapes = input_tensor_shape)
# converter.post_training_quantize = True
tflite_model = converter.convert()
with open(out_path, "wb") as f:
f.write(tflite_model)
#生成.pb .index .meta .ckpt.data文件
# freeze_graph --input_graph=/Users/jason/nsfw/flask-open-nsfw/model/nsfw-graph.pb --input_checkpoint=/Users/jason/nsfw/flask-open-nsfw/model/nsfw_model.ckpt --input_binary=true --output_graph=/Users/jason/nsfw/flask-open-nsfw/model/frozen_nsfw.pb --output_node_names=predictions
#命令行生成tflite文件(tonserflow1.3前可用)
#toco --graph_def_file=/Users/jason/nsfw/flask-open-nsfw/model/frozen_nsfw.pb --input_format=TENSORFLOW_GRAPHDEF --output_format=TFLITE --output_file=/Users/jason/nsfw/flask-open-nsfw/model/nsfw.tflite --inference_type=FLOAT --input_type=FLOAT --input_arrays=input --output_arrays=predictions input_shapes=1,224,224,3
#
# toco \
# --graph_def_file=/Users/jason/nsfw/flask-open-nsfw/model/frozen_nsfw.pb \
# --output_file=/Users/jason/nsfw/flask-open-nsfw/model/aaaa.lite \
# --input_format=TENSORFLOW_GRAPHDEF \
# --output_format=TFLITE \
# --input_shape=1,224,224,3 \
# --input_array=input \
# --output_array=predictions \
# --inference_type=FLOAT \
# --input_data_type=FLOAT
if __name__ == "__main__":
#检测加载Downloads下所有文件，逐个输出检测结果
getResultListFromDir()
#生成TFLite文件
# createTfliteFile()
# print('tensorflowVersion:',tf.__version__)
# print('npVersion:',np.__version__)