microsoft/xprophetnet-large-wiki100-cased

ProphetNet is a new pre-trained language model for sequence-to-sequence learning with a novel self-supervised objective called future n-gram prediction . The model could be trained on English headline generation as follows: input_str = "us rejects charges against its ambassador in bolivia" input_ids = tokenizer(input_str, return_tensors="pt) input_id) input labels = tokens (target_str) Label labels = labels . Labelers = labels, labels = models . Labelerizer = XLMProphetNetTokenizer.tokenizer.from_pretrained("Microsoft/xprophetnet-large-wiki100-c,