Exploring the similarity of contextual embeddings for similar sentences. Talk slides
This project supports the following languages. If you would like another, make an issue and I will add it :)
English, Spanish, Portuguese, Italian, French, German, Japanese, Chinese and Basque
I'm using Flair which supports these types of embeddings:
Word, Flair, ELMo and BERT
- Classic word embeddings
- Flair embeddings
- ELMo embeddings (quite limited as of 2020.12)
- BERT embeddings from these Huggingface 🤗 pre-trained models
- Document embeddings to obtain the embedding of the whole document/sentence