t5-weighter_cnndm-en

Model description

This model is a Classifier model based on T5-small, that predicts if a answer / question couple is considered as important fact or not (Is this answer enough relevant to appear in a plausible summary?). It is actually a component of QuestEval metric but can be used independently as it is.

How to use

from transformers import T5Tokenizer, T5ForConditionalGeneration

tokenizer = T5Tokenizer.from_pretrained("ThomasNLG/t5-weighter_cnndm-en")

model = T5ForConditionalGeneration.from_pretrained("ThomasNLG/t5-weighter_cnndm-en")

You can play with the model using the inference API, the text input format should follow this template (accordingly to the training stage of the model):

text_input = "{ANSWER} </s> {QUESTION} </s> {CONTEXT}"

Training data

The model was trained on synthetic data as described in Questeval: Summarization asks for fact-based evaluation.

Citation info

@article{scialom2021questeval,
  title={Questeval: Summarization asks for fact-based evaluation},
  author={Scialom, Thomas and Dray, Paul-Alexis and Gallinari, Patrick and Lamprier, Sylvain and Piwowarski, Benjamin and Staiano, Jacopo and Wang, Alex},
  journal={arXiv preprint arXiv:2103.12693},
  year={2021}
}
New: fine-tune this model in a few clicks by selecting AutoNLP in the "Train" menu!
Downloads last month
45
Hosted inference API
Text2Text Generation
This model can be loaded on the Inference API on-demand.
Evaluation results

Model card error

This model's model-index metadata is invalid: Schema validation error. properties must have property 'metrics'