LaBSE

Model description

Smaller Language-agnostic BERT Sentence Encoder (LaBSE) is a BERT-based model distilled from the original LaBSE model to 15 languages (from the original 109 languages) using the techniques described in the paper 'Load What You Need: Smaller Versions of Multilingual BERT' by Ukjae Jeong.

Model: HuggingFace's model hub.
Original model: TensorFlow Hub.
Distillation source: GitHub.
Conversion from TensorFlow to PyTorch: GitHub.

Usage

Using the model:

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)

To get the sentence embeddings, use the pooler output:

english_embeddings = english_outputs.pooler_output

Output for other languages:

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

For similarity between sentences, an L2-norm is recommended before calculating the similarity:

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))

Details

Details about data, training, evaluation and performance metrics are available in the original paper.

BibTeX entry and citation info

@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}
}