gss1147/ableton_god_producer_dataset
收藏资源简介:
这是一个用于训练AI成为专业Ableton Live节拍制作人的高级开放数据集,包含15,000个示例。数据集涵盖了Ableton Live 12 Suite中的多种音乐制作技术,如鼓编程与律动、贝斯音色设计(使用Serum、Massive X、Vital、Operator等)、高级音色设计与合成、混音与母带处理链、混音技术、现场表演设置以及工具调用(精确的Ableton操作)。数据集的特点是事实基础(真实的Ableton Live 12技术、精确的参数值、原生及第三方插件)、神级推理(每个示例包含深入的<think>风格推理,解释决策原因)、工具调用(精确的Ableton命令)、无占位符(具体数字、路由和工作流程)以及多种制作风格(如Trap、Drill、Lo-Fi、Future Bass、Techno、UK Garage、Phonk、Jungle等)。数据集结构包括唯一ID、类别、风格、难度、指令、输入、输出和标签等字段。
The most advanced open dataset for training AI to become a professional Ableton Live beat producer. This 15,000-example dataset trains LLMs to think and act like god-level producers in Ableton Live 12 Suite, covering drum programming & groove, bass sound design (Serum, Massive X, Vital, Operator), advanced sound design & synthesis, mixing & mastering chains, remixing techniques, live performance setup, and tool calling (exact Ableton actions). Key features include fact-based techniques, god-level reasoning, precise tool calling, no placeholders, and coverage of various production styles like Trap, Drill, Lo-Fi, Future Bass, Techno, UK Garage, Phonk, and Jungle. The dataset structure includes fields such as id, category, style, difficulty, instruction, input, output, and tags.




