TTS-AGI/majestrino-unified-detailed-captions-temporal

--- license: cc-by-4.0 task_categories: - audio-classification - automatic-speech-recognition tags: - audio - captions - temporal - majestrino pretty_name: Majestrino Unified Detailed Captions with Temporal Aspects --- # Majestrino Unified Detailed Captions with Temporal Aspects Filtered subset of [laion/majestrino-data](https://huggingface.co/datasets/laion/majestrino-data) containing only samples with `unified_detailed_caption_with_temporal_aspects`. ## Stats - **4,128,665** samples - **826** tar files (~1.1 GB each) - ~878 GB total ## Format Each tar contains paired `.flac` + `.json` files. **JSON fields:** - `caption` — the unified detailed caption with temporal aspects - `caption_type` — always `unified_detailed_caption_with_temporal_aspects` - `transcription` — speech transcription (when available, normalized from multiple source keys) - `duration` — audio duration (when available) - `characters_per_second` — CPS (when available) - `score_*` — quality scores (when available) - `emotion_whisper*` — emotion scores (when available)许可证：CC BY 4.0 任务类别： - 音频分类 - 自动语音识别 标签： - 音频 - 字幕 - 时序 - Majestrino 友好名称：Majestrino 带时序维度的统一详细字幕 # Majestrino 带时序维度的统一详细字幕 该数据集是 [laion/majestrino-data](https://huggingface.co/datasets/laion/majestrino-data) 的过滤子集，仅保留包含`unified_detailed_caption_with_temporal_aspects`字段的样本。 ## 统计信息 - **4,128,665** 条样本 - **826** 个tar文件（单个文件大小约1.1GB） - 总大小约878GB ## 数据格式 每个tar文件包含成对的`.flac`音频文件与`.json`元数据文件。 **JSON元数据字段说明：** - `caption`：带时序维度的统一详细字幕 - `caption_type`：固定为`unified_detailed_caption_with_temporal_aspects` - `transcription`：语音转录结果（若存在，会从多源键中归一化得到） - `duration`：音频时长（若存在） - `characters_per_second`：每秒字符数（CPS，若存在） - `score_*`：质量评分（若存在） - `emotion_whisper*`：情感评分（若存在）