orpheus-formatted
This dataset contains tokenized speech for Orpheus fine-tuning.
Dataset details
- Source dataset: Trelis/orpheus-ft
- Number of examples: 18
- Format: Each example contains input_ids, labels, and attention_mask in the format expected by Orpheus models.
Usage
This dataset is ready for fine-tuning Orpheus TTS models.
Creation
Created with the tokenise_speech_dataset.py script from the Trelis-Orpheus repository.