最近做深度学习需要读取hdf5文件，我读取的文件，我利用的github上分享源代码生成了hdf5文件，地址为https://github.com/tomrunia/TF_FeatureExtraction

我生成的是vgg16_features.h5

然后我读取的代码为：

#coding=utf-8
import h5py

eventNumber = 5

vgg_feature = h5py.File('vgg16_features.h5','r')
# keys=vgg_feature.keys()                            #可以查看所有的主键 
# values=vgg_feature.values()
# print(keys)
# print(values)
# print(vgg_feature.items())
filenames = vgg_feature['filenames']
vgg_16=vgg_feature['vgg_16']

for i in xrange(len(filenames)):
    arr1ev  = filenames[i]
    image_feature=vgg_16["fc7"][i]
    print(arr1ev)
    print(image_feature)

vgg_feature.close()

参考文献


[1].Datasets.http://docs.h5py.org/en/latest/high/dataset.html


[2].How to access HDF5 data from Python.https://confluence.slac.stanford.edu/display/PSDM/How+to+access+HDF5+data+from+Python