{ | |
"additional_special_tokens": [ | |
{ | |
"content": "<|ip|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|endoftext|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|startoftranscript|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|en|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|zh|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|de|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|es|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ru|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ko|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|fr|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ja|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|pt|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|tr|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|pl|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ca|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|nl|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ar|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sv|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|it|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|id|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|hi|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|fi|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|vi|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|he|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|uk|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|el|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ms|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|cs|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ro|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|da|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|hu|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ta|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|no|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|th|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ur|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|hr|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|bg|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|lt|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|la|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|mi|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ml|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|cy|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sk|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|te|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|fa|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|lv|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|bn|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sr|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|az|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sl|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|kn|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|et|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|mk|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|br|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|eu|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|is|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|hy|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ne|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|mn|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|bs|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|kk|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sq|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sw|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|gl|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|mr|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|pa|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|si|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|km|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sn|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|yo|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|so|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|af|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|oc|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ka|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|be|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|tg|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sd|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|gu|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|am|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|yi|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|lo|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|uz|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|fo|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ht|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ps|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|tk|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|nn|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|mt|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|sa|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|lb|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|my|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|bo|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|tl|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|mg|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|as|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|tt|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|haw|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ln|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ha|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|ba|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|jw|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|su|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|translate|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|transcribe|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|startoflm|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|startofprev|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|nocaptions|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
{ | |
"content": "<|notimestamps|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
} | |
], | |
"bos_token": { | |
"content": "<|endoftext|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
"eos_token": { | |
"content": "<|endoftext|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
"pad_token": { | |
"content": "<|endoftext|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
}, | |
"unk_token": { | |
"content": "<|endoftext|>", | |
"lstrip": false, | |
"normalized": false, | |
"rstrip": false, | |
"single_word": false | |
} | |
} | |