BioGPT

预训练语言模型在生物医学领域引起了越来越多的关注，受到了它们在自然语言领域取得的巨大成功的启发。在通用语言领域的两个主要预训练语言模型分支中，即BERT（及其变体）和GPT（及其变体），前者已在生物医学领域进行了广泛研究，如BioBERT和PubMedBERT。虽然它们在多种判别式下游生物医学任务上取得了巨大成功，但缺乏生成能力限制了它们的应用范围。在本文中，我们提出了BioGPT，这是一个基于大规模生物医学文献进行预训练的领域特定生成Transformer语言模型。我们在六个生物医学自然语言处理任务上评估了BioGPT，并证明我们的模型在大多数任务上优于之前的模型。特别是，我们在BC5CDR、KD-DTI和DDI端到端关系提取任务上分别获得了44.98%、38.42%和40.76%的F1得分，以及在PubMedQA上的78.2%的准确率，创造了新的记录。我们对文本生成的案例研究进一步证明了BioGPT在生成流畅的生物医学术语描述方面的优势。

您可以直接使用该模型进行文本生成。由于生成依赖于一些随机性，我们设置了一个种子以实现可重现性:

>>> from transformers import pipeline, set_seed
>>> from transformers import BioGptTokenizer, BioGptForCausalLM
>>> model = BioGptForCausalLM.from_pretrained("microsoft/biogpt")
>>> tokenizer = BioGptTokenizer.from_pretrained("microsoft/biogpt")
>>> generator = pipeline('text-generation', model=model, tokenizer=tokenizer)
>>> set_seed(42)
>>> generator("COVID-19 is", max_length=20, num_return_sequences=5, do_sample=True)
[{'generated_text': 'COVID-19 is a disease that spreads worldwide and is currently found in a growing proportion of the population'},
 {'generated_text': 'COVID-19 is one of the largest viral epidemics in the world.'},
 {'generated_text': 'COVID-19 is a common condition affecting an estimated 1.1 million people in the United States alone.'},
 {'generated_text': 'COVID-19 is a pandemic, the incidence has been increased in a manner similar to that in other'},
 {'generated_text': 'COVID-19 is transmitted via droplets, air-borne, or airborne transmission.'}]

以下是如何在PyTorch中使用该模型获取给定文本的特征:

from transformers import BioGptTokenizer, BioGptForCausalLM
tokenizer = BioGptTokenizer.from_pretrained("microsoft/biogpt")
model = BioGptForCausalLM.from_pretrained("microsoft/biogpt")
text = "Replace me by any text you'd like."
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)

波束搜索解码:

import torch
from transformers import BioGptTokenizer, BioGptForCausalLM, set_seed

tokenizer = BioGptTokenizer.from_pretrained("microsoft/biogpt")
model = BioGptForCausalLM.from_pretrained("microsoft/biogpt")

sentence = "COVID-19 is"
inputs = tokenizer(sentence, return_tensors="pt")

set_seed(42)

with torch.no_grad():
    beam_output = model.generate(**inputs,
                                min_length=100,
                                max_length=1024,
                                num_beams=5,
                                early_stopping=True
                                )
tokenizer.decode(beam_output[0], skip_special_tokens=True)
'COVID-19 is a global pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of coronavirus disease 2019 (COVID-19), which has spread to more than 200 countries and territories, including the United States (US), Canada, Australia, New Zealand, the United Kingdom (UK), and the United States of America (USA), as of March 11, 2020, with more than 800,000 confirmed cases and more than 800,000 deaths.'

引用

如果您在研究中发现BioGPT对您有用，请引用以下论文:

@article{10.1093/bib/bbac409,
    author = {Luo, Renqian and Sun, Liai and Xia, Yingce and Qin, Tao and Zhang, Sheng and Poon, Hoifung and Liu, Tie-Yan},
    title = "{BioGPT: generative pre-trained transformer for biomedical text generation and mining}",
    journal = {Briefings in Bioinformatics},
    volume = {23},
    number = {6},
    year = {2022},
    month = {09},
    abstract = "{Pre-trained language models have attracted increasing attention in the biomedical domain, inspired by their great success in the general natural language domain. Among the two main branches of pre-trained language models in the general language domain, i.e. BERT (and its variants) and GPT (and its variants), the first one has been extensively studied in the biomedical domain, such as BioBERT and PubMedBERT. While they have achieved great success on a variety of discriminative downstream biomedical tasks, the lack of generation ability constrains their application scope. In this paper, we propose BioGPT, a domain-specific generative Transformer language model pre-trained on large-scale biomedical literature. We evaluate BioGPT on six biomedical natural language processing tasks and demonstrate that our model outperforms previous models on most tasks. Especially, we get 44.98\%, 38.42\% and 40.76\% F1 score on BC5CDR, KD-DTI and DDI end-to-end relation extraction tasks, respectively, and 78.2\% accuracy on PubMedQA, creating a new record. Our case study on text generation further demonstrates the advantage of BioGPT on biomedical literature to generate fluent descriptions for biomedical terms.}",
    issn = {1477-4054},
    doi = {10.1093/bib/bbac409},
    url = {https://doi.org/10.1093/bib/bbac409},
    note = {bbac409},
    eprint = {https://academic.oup.com/bib/article-pdf/23/6/bbac409/47144271/bbac409.pdf},
}