Hugging Face's logo

RUCAIBox
/
mvp-question-answering

Text2Text Generation Transformers PyTorch English mvp text-generation Inference Endpoints
Model card Files Community
Edit model card

MVP-question-answering

The MVP-question-answering model was proposed in MVP: Multi-task Supervised Pre-training for Natural Language Generation by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.

The detailed information and instructions can be found https://github.com/RUCAIBox/MVP.

Model Description

MVP-question-answering is a prompt-based model that MVP is further equipped with prompts pre-trained using labeled question answering datasets. It is a variant (MVP+S) of our MVP MVP model. It follows a Transformer encoder-decoder architecture with layer-wise prompts.

MVP-question-answering is specially designed for question answering tasks, such as reading comprehension (SQuAD), conversational question answering (CoQA) and closed-book question-answering (Natural Questions).

Example 

>>> from transformers import MvpTokenizer, MvpForConditionalGeneration

>>> tokenizer = MvpTokenizer.from_pretrained("RUCAIBox/mvp")
>>> model = MvpForConditionalGeneration.from_pretrained("RUCAIBox/mvp-question-answering")

>>> inputs = tokenizer(
...     "Answer the following question: From which country did Angola achieve independence in 1975?",
...     return_tensors="pt",
... )
>>> generated_ids = model.generate(**inputs)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
['Portugal']

Related Models

MVP: https://huggingface.co/RUCAIBox/mvp.

Prompt-based models:

Multi-task models:

Citation 

@article{tang2022mvp,
  title={MVP: Multi-task Supervised Pre-training for Natural Language Generation},
  author={Tang, Tianyi and Li, Junyi and Zhao, Wayne Xin and Wen, Ji-Rong},
  journal={arXiv preprint arXiv:2206.12131},
  year={2022},
  url={https://arxiv.org/abs/2206.12131},
}
Downloads last month
7