MS Marco 跨编码器

本模型基于 MS Marco Passage Ranking 任务进行训练。

该模型可用于信息检索：给定一个查询，将查询与所有可能的段落进行编码（例如使用 ElasticSearch 检索）。然后按降序排列段落。更多详情请参阅 SBERT.net 检索与重排序。训练代码在此处：SBERT.net MS Marco 训练

与 SentenceTransformers 配合使用

安装 SentenceTransformers 后，使用非常简单：

from sentence_transformers import CrossEncoder

model = CrossEncoder('cross-encoder/ms-marco-MiniLM-L4-v2')
scores = model.predict([
    ("How many people live in Berlin?", "Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers."),
    ("How many people live in Berlin?", "Berlin is well known for its museums."),
])
print(scores)
# [ 9.1273365 -4.569759 ]

与 Transformers 配合使用

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model = AutoModelForSequenceClassification.from_pretrained('cross-encoder/ms-marco-MiniLM-L4-v2')
tokenizer = AutoTokenizer.from_pretrained('cross-encoder/ms-marco-MiniLM-L4-v2')

features = tokenizer(['How many people live in Berlin?', 'How many people live in Berlin?'], ['Berlin has a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers.', 'New York City is famous for the Metropolitan Museum of Art.'],  padding=True, truncation=True, return_tensors="pt")

model.eval()
with torch.no_grad():
    scores = model(**features).logits
    print(scores)

性能

下表展示了各种预训练跨编码器在 TREC Deep Learning 2019 和 MS Marco Passage Reranking 数据集上的性能表现。

注：运行时间基于 V100 GPU 计算。