UnicodeDecodeError: 'utf-8' codec can't decode byte 0xff in position 0: invalid start byte about tensorflow_input_image_by_tfrecord HOT 4 CLOSED

Thanks @yeephycho ,
I realized I had to change line of the code to

with tf.gfile.FastGFile(filename, 'rb') as f:
    image_data = f.read()

and it works now.

from tensorflow_input_image_by_tfrecord.

Hi @isalirezag .
This repo. was created a long time ago, after a few times iteration of tensorflow, I am not sure whether it still can work as months earlier.

I noticed that the last line of the terminal, the program has already write 302 images into tfrecord, so my suggestion is try to validate whether the tfrecord file contains the right images and metadata.

With regards!

from tensorflow_input_image_by_tfrecord.

it still doesnt work for me. Still showing the same error please help

from tensorflow_input_image_by_tfrecord.

The best method of reading/accessing that data is by below snippet
pd.read_csv('data.csv', encoding='latin1')
in your case try to add encoding='latin1' that's it

from tensorflow_input_image_by_tfrecord.