The model's pipeline returns an error #2

by Eduarda - opened
edited Jun 14

Just by using the default code provided in the model's page, the pipeline returns the error KeyError: 'logits'. The pipeline is loaded fine in thepipe variable but, when we test the pipeline with a string, the snippet pipe('Tinha uma [MASK] no meio do caminho.') returns an error. I have tried multiple environments, and devices and acquired the same result.

import torch
from transformers import AutoTokenizer
from transformers import AutoModelForPreTraining
from transformers import AutoModel
model = AutoModelForPreTraining.from_pretrained('neuralmind/bert-base-portuguese-cased')
tokenizer = AutoTokenizer.from_pretrained('neuralmind/bert-base-portuguese-cased', do_lower_case=False)
from transformers import pipeline
pipe = pipeline('fill-mask', model=model, tokenizer=tokenizer)
pipe('Tinha uma [MASK] no meio do caminho.')
Eduarda changed discussion status to closed
Eduarda changed discussion status to open

When the model is downloaded directly from the pipeline function, the error is not displayed and the pipe returns the results for the [MASK] normally.

pipe = pipeline('fill-mask', model='neuralmind/bert-base-portuguese-cased')

pipe("Tinha uma [MASK] no meio do caminho.")

Does any one know the reason for this?