tokenizer_config.json

{
    "model_type": "hindi_bpe",
        "vocab_size": 4477,
        "max_token_length": 64,
        "compression_ratio": 3.66,
        "special_tokens": {
            "pad_token": "",
            "unk_token": "",
            "mask_token": "",
        },
        "do_lower_case": false,
        "strip_accents": false,
        "tokenizer_class": "HindiBPE"
}