opus-mt-tc-big-ko-en / tokenizer_config.json
wosodo1891's picture
fix: separate source and target vocab
a511285
raw
history blame
365 Bytes
{
"source_lang": "ko",
"target_lang": "en",
"unk_token": "<unk>",
"eos_token": "</s>",
"pad_token": "<pad>",
"model_max_length": 512,
"sp_model_kwargs": {},
"separate_vocabs": true,
"special_tokens_map_file": null,
"name_or_path": "marian-models/opusTCv20210807-sepvoc_transformer-big_2022-07-28/ko-en",
"tokenizer_class": "MarianTokenizer"
}