wav2vec2-nepali / tokenizer_config.json
Bibek1129's picture
Upload tokenizer
f79a444 verified
{
"added_tokens_decoder": {
"1": {
"content": "१६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"2": {
"content": "३७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"3": {
"content": "५६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"4": {
"content": "२७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"5": {
"content": "५४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"6": {
"content": "८९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"7": {
"content": "२५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"8": {
"content": "४०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"9": {
"content": "६६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"10": {
"content": "९०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"11": {
"content": "४५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"12": {
"content": "९४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"13": {
"content": "४४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"14": {
"content": "५०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"15": {
"content": "२९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"16": {
"content": "८२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"17": {
"content": "८८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"18": {
"content": "६२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"19": {
"content": "१३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"21": {
"content": "१९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"22": {
"content": "५९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"23": {
"content": "६०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"24": {
"content": "४७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"25": {
"content": "५२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"26": {
"content": "३५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"28": {
"content": "९९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"29": {
"content": "२६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"30": {
"content": "७०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"31": {
"content": "३२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"32": {
"content": "७६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"33": {
"content": "४८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"34": {
"content": "९७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"35": {
"content": "२०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"36": {
"content": "९२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"38": {
"content": "१८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"40": {
"content": "८१",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"42": {
"content": "९३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"43": {
"content": "६९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"44": {
"content": "५३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"45": {
"content": "६१",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"46": {
"content": "८३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"47": {
"content": "९५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"49": {
"content": "६४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"50": {
"content": "६५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"51": {
"content": "७३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"52": {
"content": "८६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"53": {
"content": "७२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"54": {
"content": "३३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"55": {
"content": "२८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"56": {
"content": "१०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"57": {
"content": "९१",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"58": {
"content": "४२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"59": {
"content": "४६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"60": {
"content": "११",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"61": {
"content": "७५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"62": {
"content": "३०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"63": {
"content": "५८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"64": {
"content": "४३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"65": {
"content": "७८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"66": {
"content": "५१",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"67": {
"content": "५५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"68": {
"content": "२४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"70": {
"content": "८४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"71": {
"content": "७७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"72": {
"content": "६३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"73": {
"content": "८७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"75": {
"content": "६७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"76": {
"content": "२२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"78": {
"content": "१५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"79": {
"content": "१००",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"80": {
"content": "७१",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"81": {
"content": "३८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"82": {
"content": "३१",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"83": {
"content": "९८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"84": {
"content": "५७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"85": {
"content": "२१",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"86": {
"content": "४९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"87": {
"content": "१४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"88": {
"content": "८०",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"89": {
"content": "२३",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"90": {
"content": "७९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"91": {
"content": "३४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"92": {
"content": "३६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"93": {
"content": "४१",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"94": {
"content": "१७",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"95": {
"content": "१२",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"96": {
"content": "७४",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"97": {
"content": "९६",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"99": {
"content": "३९",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"100": {
"content": "८५",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"101": {
"content": "६८",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"102": {
"content": "[UNK]",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"103": {
"content": "[PAD]",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"104": {
"content": "<s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"105": {
"content": "</s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"bos_token": "<s>",
"clean_up_tokenization_spaces": true,
"do_lower_case": false,
"eos_token": "</s>",
"model_max_length": 1000000000000000019884624838656,
"pad_token": "[PAD]",
"replace_word_delimiter_char": " ",
"target_lang": null,
"tokenizer_class": "Wav2Vec2CTCTokenizer",
"unk_token": "[UNK]",
"word_delimiter_token": "|"
}