datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
LiveStreamingRiskControl
Live or Lie — Live Streaming Room Risk Assessment (May/June 2025)
This dataset contains live-streaming room interaction logs for room-level risk assessment under weak supervision. It is the official dataset for the research presented in the papers:
Outsmarting the Chameleon: Counterfactual Decoupling for Tactical OOD Shifts in Live Streaming Risk Assessment (Hugging Face Papers)
Live or Lie: Action-Aware Capsule Multiple Instance Learning for Risk Assessment in Live Streaming… See the full description on the dataset page: https://huggingface.co/datasets/ByteDance/LiveStreamingRiskControl.live-streaming-content
直播内容训练数据集
数据集简介
这个数据集包含直播记录的对话训练数据，适用于：
微调语言模型以生成直播风格的内容
构建直播内容问答系统
RAG (检索增强生成) 系统训练
数据格式
1. SFT 格式 (training_sft.jsonl)
对话式训练数据，适用于监督微调：
{
"messages": [
{"role": "system", "content": "你是一个专业的直播内容助手..."},
{"role": "user", "content": "请介绍一下..."},
{"role": "assistant", "content": "..."}
]
}
2. RAG 格式 (training_rag.jsonl)
问答对格式，包含上下文：
{
"question": "...",
"context": "...",
"answer": "...",
"metadata": {...}
}… See the full description on the dataset page: https://huggingface.co/datasets/licco123/live-streaming-content.LiveStreamingRiskControl
Dataset Card: Live Streaming Room Risk Assessment (May/June 2025)
Dataset Summary
This dataset contains live-streaming room interaction logs for room-level risk assessment under weak supervision. Each example corresponds to a single live-streaming room and is labeled as risky (> 0) or normal (= 0).
The task is designed for early detection: each room’s action sequence is truncated to the first 30 minutes, and can be structured into user–timeslot capsules for models such as… See the full description on the dataset page: https://huggingface.co/datasets/yrqiao/LiveStreamingRiskControl.Chinese_Male_Speech_Synthesis_Corpus_Live_Streaming_for_Sales
ID
King-TTS-272
Duration
4.32 hours
Language
Chinese
URL
https://dataoceanai.com/datasets/tts/chinese-male-speech-synthesis-corpus-live-streaming-for-sales/
Chinese_Female_Speech_Synthesis_Corpus_Live_Streaming_for_Sales_with_Multi_Styles
ID
King-TTS-241
Duration
8.56 hours
Speakers
100 People
Labeling Details
Pronunciation, Rhythm, Breath sounds marked with {hx}
Language
Chinese
Description
Two styles: Deep and uplifting; covers a variety of product categories including food, clothing, beauty, personal care, electronics, and home goods.
URL… See the full description on the dataset page: https://huggingface.co/datasets/DataoceanAI/Chinese_Female_Speech_Synthesis_Corpus_Live_Streaming_for_Sales_with_Multi_Styles.Chinese_Female_Speech_Synthesis_Corpus_Live_Streaming_for_Sales
ID
King-TTS-271
Duration
4.24 hours
Language
Chinese
URL
https://dataoceanai.com/datasets/tts/chinese-female-speech-synthesis-corpus-live-streaming-for-sales/
LiveStreamingRiskControl
Dataset Card: Live Streaming Room Risk Assessment (May/June 2025)
license: cc-by-4.0
pretty_name: "Live or Lie — Live Streaming Room Risk Assessment (May/June 2025)"
language:
- zh
tags:
- live-streaming
- risk-assessment
- fraud-detection
- weak-supervision
- multiple-instance-learning
- behavior-sequence
Dataset Summary
This dataset contains live-streaming room interaction logs for room-level risk assessmentunder weak supervision. Each… See the full description on the dataset page: https://huggingface.co/datasets/fyyl/LiveStreamingRiskControl.