metadata
license: mit
language:
- en
Everyday Conversations Fine-Tuning Dataset (LLaMA 3.1 - 2K)
Overview
This repository hosts the Everyday Conversations - LLaMA 3.1 - 2K dataset, a carefully curated fine-tuning dataset designed for conversational AI models. The dataset was created using the Kokoro-82M model, featuring voice samples from the af Voicepack.
Dataset Link
Features
- Voice Model: Kokoro-82M
- Voicepack: af (American, Female)
- Samples: 2,000 24kHz conversational text and audio pairs
- Language: English (American accent)