初始化项目,由ModelHub XC社区提供模型

Model: reazon-research/japanese-hubert-base-k2-rs35kh
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-08 11:40:38 +08:00
commit f2177f9d4e
8 changed files with 5490 additions and 0 deletions

41
tokenizer_config.json Normal file
View File

@@ -0,0 +1,41 @@
{
"added_tokens_decoder": {
"0": {
"content": "<blk>",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"5228": {
"content": "<unk>",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
},
"5229": {
"content": "<sos/eos>",
"lstrip": true,
"normalized": false,
"rstrip": true,
"single_word": false,
"special": false
}
},
"bos_token": "<sos/eos>",
"clean_up_tokenization_spaces": false,
"do_lower_case": false,
"eos_token": "<sos/eos>",
"extra_special_tokens": {},
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<blk>",
"processor_class": "Wav2Vec2Processor",
"replace_word_delimiter_char": " ",
"target_lang": null,
"tokenizer_class": "Wav2Vec2CTCTokenizer",
"unk_token": "<unk>",
"word_delimiter_token": "|"
}