spacy-sentence-bert 0.0.2

SpaCy models for using sentence-BERT

Sentence-BERT for spaCy

This package wraps sentence-transformers (also known as sentence-BERT) directly in spaCy. You can substitute the vectors provided in any spaCy model with vectors that have been tuned specifically for semantic similarity.

The models below are suggested for analysing sentence similarity, as the STS benchmark indicates. Keep in mind that sentence-transformers are configured with a maximum sequence length of 128. Therefore for longer texts it may be more suitable to work with other models (e.g. Universal Sentence Encoder).

Install

To install this package, you can run one of the following:

• pip install spacy_sentence_bert
• pip install git+https://github.com/MartinoMensio/spacy-sentence-bert.git

Usage

With this package installed

import spacy_sentence_bert
nlp = spacy_sentence_bert.load_model('en_bert_base_nli_cls_token')

Or if a specific model is installed (e.g. pip install https://github.com/MartinoMensio/spacy-sentence-bert/releases/download/en_bert_base_nli_cls_token-0.1.0/en_bert_base_nli_cls_token-0.1.0.tar.gz)

import spacy
nlp = spacy.load('en_bert_base_nli_cls_token')

import spacy
import spacy_sentence_bert
nlp_base = spacy.load('en')
nlp = spacy_sentence_bert.create_from(nlp_base, 'en_bert_base_nli_cls_token')
nlp.pipe_names

Full list of models https://docs.google.com/spreadsheets/d/14QplCdTCDwEmTqrn1LH4yrbKvdogK4oQvYO1K1aPR5M/edit#gid=0

sentence-BERT name	spacy model name	dimensions	language	STS benchmark
bert-base-nli-mean-tokens	en_bert_base_nli_mean_tokens	768	en	77.12
bert-base-nli-max-tokens	en_bert_base_nli_max_tokens	768	en	77.21
bert-base-nli-cls-token	en_bert_base_nli_cls_token	768	en	76.30
bert-large-nli-mean-tokens	en_bert_large_nli_mean_tokens	1024	en	79.19
bert-large-nli-max-tokens	en_bert_large_nli_max_tokens	1024	en	78.41
bert-large-nli-cls-token	en_bert_large_nli_max_tokens	1024	en	78.29
roberta-base-nli-mean-tokens	en_roberta_base_nli_mean_tokens	768	en	77.49
roberta-large-nli-mean-tokens	en_roberta_large_nli_mean_tokens	1024	en	78.69
distilbert-base-nli-mean-tokens	en_distilbert_base_nli_mean_tokens	768	en	76.97
bert-base-nli-stsb-mean-tokens	en_bert_base_nli_stsb_mean_tokens	768	en	85.14
bert-large-nli-stsb-mean-tokens	en_bert_large_nli_stsb_mean_tokens	1024	en	85.29
roberta-base-nli-stsb-mean-tokens	en_roberta_base_nli_stsb_mean_tokens	768	en	85.40
roberta-large-nli-stsb-mean-tokens	en_roberta_large_nli_stsb_mean_tokens	1024	en	86.31
distilbert-base-nli-stsb-mean-tokens	en_distilbert_base_nli_stsb_mean_tokens	768	en	84.38
distiluse-base-multilingual-cased	xx_distiluse_base_multilingual_cased	512	Arabic, Chinese, Dutch, English, French, German, Italian, Korean, Polish, Portuguese, Russian, Spanish, Turkish	80.10
xlm-r-base-en-ko-nli-ststb	xx_xlm_r_base_en_ko_nli_ststb	768	en,ko	81.47
xlm-r-large-en-ko-nli-ststb	xx_xlm_r_base_en_ko_nli_ststb	1024	en,ko	84.05

The models, when first used, download to the folder defined with TORCH_HOME in the environment variables (default ~/.cache/torch).