这应该是意见特别简单的项目,但是我却弄了好久(大哭)。用训练好的模型检查新图片我在这也卡了好久,总是提示找不到文件,最后我把所有的文件放到了一个目录下。最后终于成功了,表示很开心。
代码没有什么变化:
import numpy as np
import matplotlib.pyplot as plt
import sys
caffe_root = 'home/ubuntu/caffe/'
sys.path.append("/home/ubuntu/caffe/python")
sys.path
import caffe
caffe.set_mode_gpu()
modle_def = 'deploy.prototxt'
modle_pretrained = 'data_iter_20000.caffemodel'
net = caffe.Classifier(modle_def,modle_pretrained,
mean = np.load('mean.npy').mean(1).mean(1),
channel_swap = (2,1,0),
raw_scale =255,
image_dims=(256,256))
image_lables_filename = 'labels.txt'
lables = np.loadtxt(image_lables_filename,str,delimiter = '\t')
img = caffe.io.load_image('test/hu.jpeg')
prediction = net.predict([img])
str_gender = lables[prediction[0].argmax()]
print 'predicted class:',prediction[0].argmax()
print str_gender
plt.imshow(img)
plt.title(str_gender)
plt.show()
在这里需要把你的mean.binaryproto转为mean.npy这里也是要把程序和mean.binaryproto放到一个文件夹。
代码如下:
#!/usr/bin/python
import numpy as np
import sys
sys.path.append('/caffe/python')
import caffe
if len(sys.argv) != 3:
print "Usage: python convert_protomean.py proto.mean out.npy"
sys.exit()
blob = caffe.proto.caffe_pb2.BlobProto()
data = open( sys.argv[1] , 'rb' ).read()
blob.ParseFromString(data)
arr = np.array( caffe.io.blobproto_to_array(blob) )
out = arr[0]
np.save( sys.argv[2] , out )