transformers.js

Add start/end character offsets to token classification and return_offsets_mapping to tokenizer

#1668

Closed