drmhse/tts-rs-assets

tts-rs-assets Derived assets for tts-rs: per-stage reference activations that the Rust ports are gated against, plus two small CosyVoice artifacts that cannot be produced without the upstream python package. No model weights are here. You download those from their own repositories; see docs/setup.md in tts-rs. What is in here, and why it is not in the git repo file size what it is cosyvoice/rand_noise.safetensors 4.6 MB The CFM decoder's fixed noise… See the full description on the dataset page: https://huggingface.co/datasets/drmhse/tts-rs-assets.