Oh the example has one pairs of [] that is not necessary, will fix. It should be:
encoding = tokenizer([prompt, prompt, prompt], [choice0, choice1, choice2], return_tensors='tf', padding=True)
1 Like