docs.rs failed to build rlx-tiny-tts-0.2.9
Please check the build logs for more information.
See Builds for ideas on how to fix a failed build, or Metadata for how to configure docs.rs builds.
If you believe this is docs.rs' fault, open an issue.
Please check the build logs for more information.
See Builds for ideas on how to fix a failed build, or Metadata for how to configure docs.rs builds.
If you believe this is docs.rs' fault, open an issue.
TinyTTS English text-to-speech for RLX.
TinyTTS (https://github.com/tronghieuit/tiny-tts) is a MeloTTS / VITS2-style
model exported as four ONNX subgraphs — text_encoder, duration_predictor,
flow and decoder — with a small NumPy "glue" stage (monotonic alignment +
latent sampling) connecting them. RLX imports each ONNX graph into the rlx-ir
HIR and runs it on any backend (CPU / Metal / MLX / CUDA / ROCm / wgpu); the
glue stage is reimplemented in Rust (see [glue]).