When I run image_dataset_to_tfrecords.py ,it Report errors"KeyError: 'training\\n01443537'"

[IreneTa]

When I run image_dataset_to_tfrecords.py ,it Report errors"KeyError: 'training\n01443537'" I have no idea why,Does anyone know why?

[lxlanyu]

in image_dataset_to_tfrecords.py: for dir_path, _, files in subdirs: print('dirpath:', dir_path) # dir_name = dir_path.split('/')[-1] # dir_name: training\n03902125 dir_name = dir_path.split('\')[-1] # dir_name: n03902125 print('dir_name:', dir_name) for file_name in files: # print('file_name', file_name) # file_name n03902125_0.JPEG image_metadata = [dir_name, os.path.join(dir_name, file_name)] # print('image_metadata:', image_metadata, type(image_metadata)) # ['training\n12267677', 'training\n12267677\n12267677_99.JPEG'] metadata.append(image_metadata) I also encountered this problem and found that my dir_name is wrong. It should be n03902125 instead of training\n03902125. So you should modify the dir_name = dir_path.split('/')[-1] statement so that its result is n03902125.