LaBSE

模型描述

Smaller Language-agnostic BERT Sentence Encoder (LaBSE) 是基于BERT的模型，通过使用论文中所描述的技术，从原始的109种语言中提取出15种语言，并对其进行了压缩。

• 模型: HuggingFace's model hub 。
• 原始模型: TensorFlow Hub 。
• 压缩来源: GitHub 。
• 从TensorFlow转换为PyTorch: GitHub 。

使用方法

使用该模型：

import torch
from transformers import BertModel, BertTokenizerFast


tokenizer = BertTokenizerFast.from_pretrained("setu4993/smaller-LaBSE")
model = BertModel.from_pretrained("setu4993/smaller-LaBSE")
model = model.eval()

english_sentences = [
    "dog",
    "Puppies are nice.",
    "I enjoy taking long walks along the beach with my dog.",
]
english_inputs = tokenizer(english_sentences, return_tensors="pt", padding=True)

with torch.no_grad():
    english_outputs = model(**english_inputs)

要获取句子嵌入，使用汇聚器输出：

english_embeddings = english_outputs.pooler_output

其他语言的输出：

italian_sentences = [
    "cane",
    "I cuccioli sono carini.",
    "Mi piace fare lunghe passeggiate lungo la spiaggia con il mio cane.",
]
japanese_sentences = ["犬", "子犬はいいです", "私は犬と一緒にビーチを散歩するのが好きです"]
italian_inputs = tokenizer(italian_sentences, return_tensors="pt", padding=True)
japanese_inputs = tokenizer(japanese_sentences, return_tensors="pt", padding=True)

with torch.no_grad():
    italian_outputs = model(**italian_inputs)
    japanese_outputs = model(**japanese_inputs)

italian_embeddings = italian_outputs.pooler_output
japanese_embeddings = japanese_outputs.pooler_output

对于句子之间的相似度比较，在计算相似度之前，建议进行L2范数归一化：

import torch.nn.functional as F


def similarity(embeddings_1, embeddings_2):
    normalized_embeddings_1 = F.normalize(embeddings_1, p=2)
    normalized_embeddings_2 = F.normalize(embeddings_2, p=2)
    return torch.matmul(
        normalized_embeddings_1, normalized_embeddings_2.transpose(0, 1)
    )


print(similarity(english_embeddings, italian_embeddings))
print(similarity(english_embeddings, japanese_embeddings))
print(similarity(italian_embeddings, japanese_embeddings))

详细信息

关于数据、训练、评估和性能指标的详细信息，请参阅 original paper 。

BibTeX条目和引用信息

@misc{feng2020languageagnostic,
      title={Language-agnostic BERT Sentence Embedding},
      author={Fangxiaoyu Feng and Yinfei Yang and Daniel Cer and Naveen Arivazhagan and Wei Wang},
      year={2020},
      eprint={2007.01852},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}