模型:
microsoft/prophetnet-large-uncased-squad-qg
prophetnet-large-uncased-squad-qg Fine-tuned weights(converted from original fairseq version repo ) for ProphetNet on question generation SQuAD 1.1. ProphetNet is a new pre-trained language model for sequence-to-sequence learning with a novel self-supervised objective called future n-gram prediction. ProphetNet is able to predict more future tokens with a n-stream decoder. The original implementation is Fairseq version at github repo .
from transformers import ProphetNetTokenizer, ProphetNetForConditionalGeneration, ProphetNetConfig model = ProphetNetForConditionalGeneration.from_pretrained('microsoft/prophetnet-large-uncased-squad-qg') tokenizer = ProphetNetTokenizer.from_pretrained('microsoft/prophetnet-large-uncased-squad-qg') FACT_TO_GENERATE_QUESTION_FROM = ""Bill Gates [SEP] Microsoft was founded by Bill Gates and Paul Allen on April 4, 1975." inputs = tokenizer([FACT_TO_GENERATE_QUESTION_FROM], return_tensors='pt') # Generate Summary question_ids = model.generate(inputs['input_ids'], num_beams=5, early_stopping=True) tokenizer.batch_decode(question_ids, skip_special_tokens=True) # should give: 'along with paul allen, who founded microsoft?'
@article{yan2020prophetnet, title={Prophetnet: Predicting future n-gram for sequence-to-sequence pre-training}, author={Yan, Yu and Qi, Weizhen and Gong, Yeyun and Liu, Dayiheng and Duan, Nan and Chen, Jiusheng and Zhang, Ruofei and Zhou, Ming}, journal={arXiv preprint arXiv:2001.04063}, year={2020} }