chunked/nltk_chunking/model_doc_xglm/chunk_2.txt

While these models are known to be able to jointly represent many different languages, 
their training data is dominated by English, potentially limiting their cross-lingual generalization.