translate-windy-core / tokenizer_config.json
sneakyfree's picture
Dr. F: CT2 INT8 + minimal LoRA fog-of-mirror; FLORES-200 verified no regression
968c5b7 verified
Raw
History Blame Contribute Delete
333 Bytes
{
"src_lang": null,
"tgt_lang": null,
"bos_token": "<s>",
"eos_token": "</s>",
"sep_token": "</s>",
"unk_token": "<unk>",
"pad_token": "<pad>",
"special_tokens_map_file": "m2m_100_1.2B_v2/special_tokens_map.json",
"tokenizer_file": null,
"name_or_path": "m2m_100_1.2B_v2",
"tokenizer_class": "M2M100Tokenizer"
}