whisper-large-v2-no-digits-norm-punct / tokenizer_config.json
Den4ikAI's picture
Upload folder using huggingface_hub
f9bcb81 verified
{
"add_bos_token": false,
"add_prefix_space": false,
"added_tokens_decoder": {
"49639": {
"content": "<|endoftext|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49640": {
"content": "<|startoftranscript|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49641": {
"content": "<|en|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49642": {
"content": "<|zh|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49643": {
"content": "<|de|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49644": {
"content": "<|es|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49645": {
"content": "<|ru|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49646": {
"content": "<|ko|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49647": {
"content": "<|fr|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49648": {
"content": "<|ja|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49649": {
"content": "<|pt|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49650": {
"content": "<|tr|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49651": {
"content": "<|pl|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49652": {
"content": "<|ca|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49653": {
"content": "<|nl|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49654": {
"content": "<|ar|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49655": {
"content": "<|sv|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49656": {
"content": "<|it|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49657": {
"content": "<|id|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49658": {
"content": "<|hi|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49659": {
"content": "<|fi|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49660": {
"content": "<|vi|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49661": {
"content": "<|he|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49662": {
"content": "<|uk|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49663": {
"content": "<|el|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49664": {
"content": "<|ms|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49665": {
"content": "<|cs|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49666": {
"content": "<|ro|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49667": {
"content": "<|da|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49668": {
"content": "<|hu|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49669": {
"content": "<|ta|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49670": {
"content": "<|no|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49671": {
"content": "<|th|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49672": {
"content": "<|ur|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49673": {
"content": "<|hr|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49674": {
"content": "<|bg|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49675": {
"content": "<|lt|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49676": {
"content": "<|la|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49677": {
"content": "<|mi|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49678": {
"content": "<|ml|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49679": {
"content": "<|cy|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49680": {
"content": "<|sk|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49681": {
"content": "<|te|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49682": {
"content": "<|fa|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49683": {
"content": "<|lv|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49684": {
"content": "<|bn|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49685": {
"content": "<|sr|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49686": {
"content": "<|az|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49687": {
"content": "<|sl|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49688": {
"content": "<|kn|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49689": {
"content": "<|et|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49690": {
"content": "<|mk|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49691": {
"content": "<|br|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49692": {
"content": "<|eu|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49693": {
"content": "<|is|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49694": {
"content": "<|hy|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49695": {
"content": "<|ne|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49696": {
"content": "<|mn|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49697": {
"content": "<|bs|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49698": {
"content": "<|kk|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49699": {
"content": "<|sq|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49700": {
"content": "<|sw|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49701": {
"content": "<|gl|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49702": {
"content": "<|mr|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49703": {
"content": "<|pa|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49704": {
"content": "<|si|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49705": {
"content": "<|km|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49706": {
"content": "<|sn|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49707": {
"content": "<|yo|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49708": {
"content": "<|so|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49709": {
"content": "<|af|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49710": {
"content": "<|oc|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49711": {
"content": "<|ka|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49712": {
"content": "<|be|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49713": {
"content": "<|tg|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49714": {
"content": "<|sd|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49715": {
"content": "<|gu|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49716": {
"content": "<|am|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49717": {
"content": "<|yi|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49718": {
"content": "<|lo|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49719": {
"content": "<|uz|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49720": {
"content": "<|fo|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49721": {
"content": "<|ht|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49722": {
"content": "<|ps|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49723": {
"content": "<|tk|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49724": {
"content": "<|nn|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49725": {
"content": "<|mt|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49726": {
"content": "<|sa|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49727": {
"content": "<|lb|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49728": {
"content": "<|my|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49729": {
"content": "<|bo|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49730": {
"content": "<|tl|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49731": {
"content": "<|mg|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49732": {
"content": "<|as|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49733": {
"content": "<|tt|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49734": {
"content": "<|haw|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49735": {
"content": "<|ln|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49736": {
"content": "<|ha|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49737": {
"content": "<|ba|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49738": {
"content": "<|jw|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49739": {
"content": "<|su|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49740": {
"content": "<|translate|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49741": {
"content": "<|transcribe|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49742": {
"content": "<|startoflm|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49743": {
"content": "<|startofprev|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49744": {
"content": "<|nocaptions|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"49745": {
"content": "<|notimestamps|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"<|endoftext|>",
"<|startoftranscript|>",
"<|en|>",
"<|zh|>",
"<|de|>",
"<|es|>",
"<|ru|>",
"<|ko|>",
"<|fr|>",
"<|ja|>",
"<|pt|>",
"<|tr|>",
"<|pl|>",
"<|ca|>",
"<|nl|>",
"<|ar|>",
"<|sv|>",
"<|it|>",
"<|id|>",
"<|hi|>",
"<|fi|>",
"<|vi|>",
"<|he|>",
"<|uk|>",
"<|el|>",
"<|ms|>",
"<|cs|>",
"<|ro|>",
"<|da|>",
"<|hu|>",
"<|ta|>",
"<|no|>",
"<|th|>",
"<|ur|>",
"<|hr|>",
"<|bg|>",
"<|lt|>",
"<|la|>",
"<|mi|>",
"<|ml|>",
"<|cy|>",
"<|sk|>",
"<|te|>",
"<|fa|>",
"<|lv|>",
"<|bn|>",
"<|sr|>",
"<|az|>",
"<|sl|>",
"<|kn|>",
"<|et|>",
"<|mk|>",
"<|br|>",
"<|eu|>",
"<|is|>",
"<|hy|>",
"<|ne|>",
"<|mn|>",
"<|bs|>",
"<|kk|>",
"<|sq|>",
"<|sw|>",
"<|gl|>",
"<|mr|>",
"<|pa|>",
"<|si|>",
"<|km|>",
"<|sn|>",
"<|yo|>",
"<|so|>",
"<|af|>",
"<|oc|>",
"<|ka|>",
"<|be|>",
"<|tg|>",
"<|sd|>",
"<|gu|>",
"<|am|>",
"<|yi|>",
"<|lo|>",
"<|uz|>",
"<|fo|>",
"<|ht|>",
"<|ps|>",
"<|tk|>",
"<|nn|>",
"<|mt|>",
"<|sa|>",
"<|lb|>",
"<|my|>",
"<|bo|>",
"<|tl|>",
"<|mg|>",
"<|as|>",
"<|tt|>",
"<|haw|>",
"<|ln|>",
"<|ha|>",
"<|ba|>",
"<|jw|>",
"<|su|>",
"<|translate|>",
"<|transcribe|>",
"<|startoflm|>",
"<|startofprev|>",
"<|nocaptions|>",
"<|notimestamps|>"
],
"bos_token": "<|endoftext|>",
"clean_up_tokenization_spaces": true,
"eos_token": "<|endoftext|>",
"errors": "replace",
"model_max_length": 1024,
"pad_token": "<|endoftext|>",
"processor_class": "WhisperProcessor",
"return_attention_mask": false,
"tokenizer_class": "WhisperTokenizer",
"unk_token": "<|endoftext|>"
}