Create a component that takes the embedding vector of each token and writes it to file.
This makes so sense for static embeddings, of course, but is meant to be used with contextualized embeddings in order to collect the different embeddings for specific words.
Create a component that takes the embedding vector of each token and writes it to file. This makes so sense for static embeddings, of course, but is meant to be used with contextualized embeddings in order to collect the different embeddings for specific words.