TypeError: a bytes-like object is required, not 'dict'

[surabhshahmobiliya]

I am trying to implement ../keras-text-classification/20_news_group_classification.py

However its showing me the error as shown in the image

Libraries installed

keras==2.2.4
tensorflow==1.0.0
Python 3.6.6

[hussamsayeed]

Did you resolved this?

[DiegoAnas]

I had the same error, but I fixed it by converting the list, in your case train_posts, which was a list of bytes into a list of str. You can do that easily saving them all into another list with .decode('utf-8')

[DullPointerException]

DiegoAnas' solution worked for me.

Had this same issue when evecuting :
tokenizer.fit_on_texts(training_sentences)

Changed it to :
tokenizer.fit_on_texts([x.decode('utf-8') for x in training_sentences] )

It worked fine now

[wikazp]

Having the same error. Can anyone help me out on this?