{ "add_bos_token": false, "add_prefix_space": false, "added_tokens_decoder": { "50257": { "content": "<|endoftext|>", "lstrip": false, "normalized": true, "rstrip": false, "single_word": false, "special": true }, "50258": { "content": "<|startoftranscript|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50259": { "content": "<|en|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50260": { "content": "<|zh|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50261": { "content": "<|de|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50262": { "content": "<|es|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50263": { "content": "<|ru|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50264": { "content": "<|ko|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50265": { "content": "<|fr|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50266": { "content": "<|ja|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50267": { "content": "<|pt|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50268": { "content": "<|tr|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50269": { "content": "<|pl|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50270": { "content": "<|ca|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50271": { "content": "<|nl|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50272": { "content": "<|ar|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50273": { "content": "<|sv|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50274": { "content": "<|it|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50275": { "content": "<|id|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50276": { "content": "<|hi|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50277": { "content": "<|fi|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50278": { "content": "<|vi|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50279": { "content": "<|he|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50280": { "content": "<|uk|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50281": { "content": "<|el|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50282": { "content": "<|ms|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50283": { "content": "<|cs|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50284": { "content": "<|ro|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50285": { "content": "<|da|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50286": { "content": "<|hu|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50287": { "content": "<|ta|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50288": { "content": "<|no|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50289": { "content": "<|th|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50290": { "content": "<|ur|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50291": { "content": "<|hr|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50292": { "content": "<|bg|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50293": { "content": "<|lt|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50294": { "content": "<|la|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50295": { "content": "<|mi|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50296": { "content": "<|ml|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50297": { "content": "<|cy|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50298": { "content": "<|sk|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50299": { "content": "<|te|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50300": { "content": "<|fa|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50301": { "content": "<|lv|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50302": { "content": "<|bn|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50303": { "content": "<|sr|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50304": { "content": "<|az|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50305": { "content": "<|sl|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50306": { "content": "<|kn|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50307": { "content": "<|et|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50308": { "content": "<|mk|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50309": { "content": "<|br|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50310": { "content": "<|eu|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50311": { "content": "<|is|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50312": { "content": "<|hy|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50313": { "content": "<|ne|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50314": { "content": "<|mn|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50315": { "content": "<|bs|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50316": { "content": "<|kk|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50317": { "content": "<|sq|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50318": { "content": "<|sw|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50319": { "content": "<|gl|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50320": { "content": "<|mr|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50321": { "content": "<|pa|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50322": { "content": "<|si|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50323": { "content": "<|km|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50324": { "content": "<|sn|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50325": { "content": "<|yo|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50326": { "content": "<|so|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50327": { "content": "<|af|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50328": { "content": "<|oc|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50329": { "content": "<|ka|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50330": { "content": "<|be|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50331": { "content": "<|tg|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50332": { "content": "<|sd|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50333": { "content": "<|gu|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50334": { "content": "<|am|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50335": { "content": "<|yi|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50336": { "content": "<|lo|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50337": { "content": "<|uz|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50338": { "content": "<|fo|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50339": { "content": "<|ht|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50340": { "content": "<|ps|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50341": { "content": "<|tk|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50342": { "content": "<|nn|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50343": { "content": "<|mt|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50344": { "content": "<|sa|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50345": { "content": "<|lb|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50346": { "content": "<|my|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50347": { "content": "<|bo|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50348": { "content": "<|tl|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50349": { "content": "<|mg|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50350": { "content": "<|as|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50351": { "content": "<|tt|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50352": { "content": "<|haw|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50353": { "content": "<|ln|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50354": { "content": "<|ha|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50355": { "content": "<|ba|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50356": { "content": "<|jw|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50357": { "content": "<|su|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50358": { "content": "<|translate|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50359": { "content": "<|transcribe|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50360": { "content": "<|startoflm|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50361": { "content": "<|startofprev|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50362": { "content": "<|nocaptions|>", "lstrip": true, "normalized": false, "rstrip": true, "single_word": false, "special": true }, "50363": { "content": "<|notimestamps|>", "lstrip": true, "normalized": false, "rstrip": true, "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, "no_timestamps": true, "pad_token": "<|endoftext|>", "processor_class": "WhisperProcessor", "return_attention_mask": false, "tokenizer_class": "WhisperTokenizer", "unk_token": "<|endoftext|>" }