[SAMPLE] Large Language Model (LLM) Data | 800,000 SFX Professional Sound Effects | Human Metadata

Our audio dataset stands out from the rest and is ideal for Large Language Model (LLM) Data use cases. We boast the most widely used sound library globally, featuring nearly 800,000 sounds employed by top names like Disney, BBC, Pixar, Apple, Ogilvy, Saatchi & Saatchi, HBO, and Activision. Our sounds are recorded by professionals responsible for the audio in films like Mad Max: Fury Road, The Revenant, The Triangle of Sadness, and Dunkirk. Each audio file includes meticulously crafted metadata: a brief description, categorized listings, and keyword/tags. The dataset spans all categories, including Ambiances, Animals, Foley, Transport, Weapons, Industrial, Sports, and more. Additionally, we provide access to 30,000 music tracks with stems, all pre-cleared for machine learning and AI use.

搜集汇总

背景与挑战

背景概述

该数据集专为大型语言模型优化，包含约80万个由专业人士录制的音效，覆盖环境、动物、交通等全类别，每个文件均附有描述、分类和标签等精细元数据。此外，还提供3万首已获机器学习使用许可的音乐曲目，广泛应用于迪士尼、BBC等顶级机构。