eureka1500/CaptionStew10M-Qwen3Omni

CaptionStew是一个大规模音频字幕数据集，包含912,931个训练示例，数据以英语呈现。该数据集专注于音频内容与文本字幕的关联，可能用于训练或评估音频字幕生成模型，特别是与Qwen3-Omni-Captioner技术相关。数据文件为JSONL格式（CaptionStew_10M_30s_00.jsonl），适用于自然语言处理和音频处理任务。CaptionStew is a large-scale audio caption dataset containing 912,931 training examples, presented in English. The dataset focuses on the association between audio content and text captions, likely intended for training or evaluating audio caption generation models, particularly related to Qwen3-Omni-Captioner technology. The data is in JSONL format (CaptionStew_10M_30s_00.jsonl) and suitable for natural language processing and audio processing tasks.