tokenizer.json

{
    "module": "keras_nlp.src.models.whisper.whisper_tokenizer",
    "class_name": "WhisperTokenizer",
    "config": {
        "name": "whisper_tokenizer",
        "trainable": true,
        "dtype": "int32",
        "sequence_length": null,
        "add_prefix_space": false,
        "special_tokens": {
            "<|startoftranscript|>": 50257,
            "<|endoftext|>": 50256,
            "<|notimestamps|>": 50362,
            "<|translate|>": 50358,
            "<|transcribe|>": 50357
        },
        "language_tokens": null
    },
    "registered_name": "keras_nlp>WhisperTokenizer",
    "assets": [
        "assets/tokenizer/merges.txt",
        "assets/tokenizer/vocabulary.json"
    ],
    "weights": null
}