Upload processor

This commit is contained in:
Yoach Lacombe 2024-10-01 08:14:50 +00:00 committed by system
parent 57d207fdeb
commit 7e2743d4d1
No known key found for this signature in database
GPG Key ID: 6A528E38E0733467
3 changed files with 1752 additions and 2 deletions

1742
normalizer.json Normal file

File diff suppressed because it is too large Load Diff

@ -122,6 +122,13 @@
"rstrip": false, "rstrip": false,
"single_word": false "single_word": false
}, },
"pad_token": {
"content": "<|endoftext|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false
},
"unk_token": { "unk_token": {
"content": "<|endoftext|>", "content": "<|endoftext|>",
"lstrip": false, "lstrip": false,

@ -12984,12 +12984,13 @@
"<|notimestamps|>" "<|notimestamps|>"
], ],
"bos_token": "<|endoftext|>", "bos_token": "<|endoftext|>",
"clean_up_tokenization_spaces": false, "clean_up_tokenization_spaces": true,
"eos_token": "<|endoftext|>", "eos_token": "<|endoftext|>",
"errors": "replace", "errors": "replace",
"model_max_length": 1000000000000000019884624838656, "model_max_length": 1000000000000000019884624838656,
"pad_token": null, "pad_token": "<|endoftext|>",
"processor_class": "WhisperProcessor", "processor_class": "WhisperProcessor",
"tokenizer_class": "WhisperTokenizer", "tokenizer_class": "WhisperTokenizer",
"trust_remote_code": false,
"unk_token": "<|endoftext|>" "unk_token": "<|endoftext|>"
} }