Add a LTP WordSplitter
into AllenNLP's tokenizers.
{
"dataset_reader": {
// ... ...
"tokenizer": {
"word_splitter": {
"type": "ltp_remote",
"url": "http://10.1.1.174:12345/ltp"
}
},
// ... ...
},
// ... ...
}
allennlp train --include-package allennlp_wordsplitter_ltp -s /your/output/dir /your/training/config/file