1、AttributeError: ‘str’ object has no attribute ‘decode’
File “E:\bianchengruanjian\python 3.6\python 3.6.6\lib\site-packages\keras\engine\saving.py”, line 1145, in load_weights_from_hdf5_group
original_keras_version = f.attrs['keras_version'].decode('utf8')
AttributeError: ‘str’ object has no attribute ‘decode’
解决:卸载原来的h5py,下载低版本
如:pip install h5py==2.10 -i https://pypi.tuna.tsinghua.edu.cn/simple/
2、ValueError: Object arrays cannot be loaded when allow_pickle=False
File “E:\bianchengruanjian\python 3.6\python 3.6.6\lib\site-packages\numpy\lib\format.py”, line 727, in read_array
raise ValueError("Object arrays cannot be loaded when "
ValueError: Object arrays cannot be loaded when allow_pickle=False
解决办法:
a:下载1.16.2版本numpy
3、AttributeError: module keras.backend has no attribute image_dim_ordering
File “F:\keti\mubiao\ziliao\zero shot\ZSD_Release-master\keras_frcnn\resnet.py”, line 159, in nn_base
if K.image_dim_ordering() == 'th':
AttributeError: module ‘keras.backend’ has no attribute ‘image_dim_ordering’
解决:代码if K.image_dim_ordering() == 'th'
是旧版本keras,新版本将代码改为 K.image_data_format() == "channels_first"
4、AttributeError: module ‘keras.backend’ has no attribute ‘image_dim_ordering’
File “F:\keti\mubiao\ziliao\zero shot\ZSD_Release-master\keras_frcnn\resnet.py”, line 173, in nn_base
if K.image_dim_ordering() == 'tf':
AttributeError: module ‘keras.backend’ has no attribute ‘image_dim_ordering’
解决:将代码if K.image_dim_ordering() == 'tf'
改为K.image_data_format()== ‘channels_last’