matheusfpinto/orpheus-ptbr-snac-8192

Orpheus PT-BR SNAC 8192数据集来源于`opedromartins/ASR-datasets-ptbr`，经过文本分词和音频编码处理。数据集的序列格式包含文本标记ID和SNAC音频ID，序列长度为8192个token。数据字段包括`input_ids`、`attention_mask`和`metadata`，其中`metadata`包含原始数据集、配置、分割、音频长度和文本信息。The Orpheus PT-BR SNAC 8192 dataset is sourced from `opedromartins/ASR-datasets-ptbr` and processed with text tokenization and audio coding. The sequence format includes text token IDs and SNAC audio IDs, with a sequence length of 8192 tokens. The data fields consist of `input_ids`, `attention_mask`, and `metadata`, where `metadata` contains the original dataset, configuration, split, audio length, and text information.