All models - Vector Hub by Vector AI

Albert - A Lite Bert

We propose a novel method to reduce the memory consumption of BERT, and show that it improves the scalability of BERT models.

Release date: 2019-09-26

0 Stars pip install vectorhub[encoders-text-tfhub]

BERT - Bidirectional Encoder Representations

Unlike recent language representation models, BERT is designed to pre-train deep bidirectional representations from unlabeled text by jointly conditioning on both left and right context in all layers.

Release date: 2018-10-11

0 Stars pip install vectorhub[encoders-text-tfhub]

Experts BERT - Collection of BERT experts fine-tuned on different datasets.

For example, one can use a BERT model that was trained on text from a similar domain or by use a BERT model that was trained for a similar task.

Release date: 2021-02-15

0 Stars pip install vectorhub[encoders-text-tfhub]

Elmo (Embeddings From Language Models)

ELMo is a deep, character-based, bidirectional language model that learns to embed words in a way that captures their context.

Release date: 2020-07-03

0 Stars pip install vectorhub[encoders-text-tfhub]

LaBSE - Language-agnostic BERT Sentence Embedding

We present a multilingual BERT embedding model, called LaBSE, that produces language-agnostic cross-lingual sentence embeddings for 109 languages.

Release date: 2020-07-03

0 Stars pip install vectorhub[encoders-text-tfhub]

USE - Universal Sentence Encoder

Greater-than-word length text encoder for semantic search.

Release date: 2018-03-29

0 Stars pip install vectorhub[encoders-text-tfhub]

USE Multi - Universal Sentence Encoder Multilingual

Supports 16 languages (Arabic, Chinese-simplified, Chinese-traditional, English, French, German, Italian, Japanese, Korean, Dutch, Polish, Portuguese, Spanish, Thai, Turkish, Russian) text encoder.

Release date: 2018-03-29

0 Stars pip install vectorhub[encoders-text-tfhub]

USE Lite - Universal Sentence Encoder Lite

This lite version is good for use cases when your computation resource is limited.

Release date: 2018-03-29

0 Stars pip install vectorhub[encoders-text-tfhub]

Viewing all models

Albert - A Lite Bert

BERT - Bidirectional Encoder Representations

Experts BERT - Collection of BERT experts fine-tuned on different datasets.

Elmo (Embeddings From Language Models)

LaBSE - Language-agnostic BERT Sentence Embedding

USE - Universal Sentence Encoder

USE Multi - Universal Sentence Encoder Multilingual

USE Lite - Universal Sentence Encoder Lite