freezer232/bella-tokenized
收藏官方服务:
资源简介:
bella-tokenized数据集包含60个经过分词处理的语音样本,适用于Orpheus语音合成模型的微调。每个样本包括input_ids、labels和attention_mask,符合Orpheus模型输入数据的格式要求。
The bella-tokenized dataset contains 60 tokenized speech samples, suitable for fine-tuning Orpheus speech synthesis models. Each sample includes input_ids, labels, and attention_mask, which meet the format requirements of the Orpheus model input data.
提供机构:
freezer232


