pipeline_tag: text-to-speech

Christina-TTS-I-Beta

A TTS model trained on 50% of the Christina-TTS-I dataset for testing purposes.

Details

• Trained: trained in under 15 minutes on partial dataset (50% portion)
• Hardware: Single NVIDIA H100 NVL GPU

Audio Sample

Listen to a sample generated by the model:

Warnings

• Limited training: Inconsistent quality, noise and artifacts.
• Some mispronunciations were noticed.