Unrecognized configuration class in mT5-small-finetuned-tydiqa-for-xqa


I tried to run the multilingual question-answering with the mt5 model at https://huggingface.co/mrm8488/mT5-small-finetuned-tydiqa-for-xqa. Unfortunately I couldn’t and the following message appears:

Unrecognized configuration class <class ‘transformers.models.t5.configuration_t5.T5Config’> for this kind of AutoModel: AutoModelForCausalLM.
Model type should be one of CamembertConfig, XLMRobertaConfig, RobertaConfig, BertConfig, OpenAIGPTConfig, GPT2Config, TransfoXLConfig, XLNetConfig, XLMConfig, CTRLConfig, ReformerConfig, BertGenerationConfig, XLMProphetNetConfig, ProphetNetConfig.

How can this be solved?


Hi there, could you post the code snippet the raised this error?


This is the code:

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
tokenizer = AutoTokenizer.from_pretrained("mrm8488/mT5-small-finetuned-tydiqa-for-xqa")
model = AutoModelForCausalLM.from_pretrained("mrm8488/mT5-small-finetuned-tydiqa-for-xqa").to(device)

def get_response(question, context, max_length=32):
  input_text = 'question: %s  context: %s' % (question, context)
  features = tokenizer([input_text], return_tensors='pt')

  output = model.generate(input_ids=features['input_ids'].to(device), 

  return tokenizer.decode(output[0])

# Some examples in different languages

context = 'HuggingFace won the best Demo paper at EMNLP2020.'
question = 'What won HuggingFace?'
get_response(question, context)

context = 'HuggingFace ganó la mejor demostración con su paper en la EMNLP2020.'
question = 'Qué ganó HuggingFace?'
get_response(question, context)

context = 'HuggingFace выиграл лучшую демонстрационную работу на EMNLP2020.'
question = 'Что победило в HuggingFace?'
get_response(question, context)

It is the same you can find in https://huggingface.co/mrm8488/mT5-small-finetuned-tydiqa-for-xqa


the issue is MT5 is a seq2seq model and seq2seq models should be loaded using
AutoModelForConditionalGeneration or in this case directly using MT5ForConditionalGeneration class.

Also cc @mrm8488

Good morning

Thank you very much. Now the error is not raised

Thanks again

1 Like

I will fix it ASAP