xlm-mlm-17-1280 / tokenizer_config.json
Xenova's picture
Xenova HF staff
Upload folder using huggingface_hub
50d74ba
{
"additional_special_tokens": [
"<special0>",
"<special1>",
"<special2>",
"<special3>",
"<special4>",
"<special5>",
"<special6>",
"<special7>",
"<special8>",
"<special9>"
],
"bos_token": "<s>",
"clean_up_tokenization_spaces": true,
"cls_token": "</s>",
"do_lowercase_and_remove_accent": false,
"id2lang": {
"0": "ar",
"1": "de",
"10": "pl",
"11": "pt",
"12": "ru",
"13": "sv",
"14": "tr",
"15": "vi",
"16": "zh",
"2": "en",
"3": "es",
"4": "fr",
"5": "hi",
"6": "it",
"7": "ja",
"8": "ko",
"9": "nl"
},
"lang2id": {
"ar": 0,
"de": 1,
"en": 2,
"es": 3,
"fr": 4,
"hi": 5,
"it": 6,
"ja": 7,
"ko": 8,
"nl": 9,
"pl": 10,
"pt": 11,
"ru": 12,
"sv": 13,
"tr": 14,
"vi": 15,
"zh": 16
},
"mask_token": "<special1>",
"model_max_length": 512,
"pad_token": "<pad>",
"sep_token": "</s>",
"tokenizer_class": "XLMTokenizer",
"unk_token": "<unk>"
}