MLACookie 

MossTtsSharp backend abstraction layer and auto-detection — defines BackendProvider for platform-specific ONNX Runtime session creation.

MossTtsSharp C# inference library — text-to-speech synthesis using ONNX Runtime. Supports CPU, DirectML (Windows), and CUDA (Linux) backends.

MossTtsSharp CUDA backend — adds CUDA ONNX Runtime for GPU-accelerated inference on NVIDIA GPUs.

MossTtsSharp DirectML backend — adds DirectML ONNX Runtime for GPU-accelerated inference on Windows.

MossTtsSharp Cpu backend — adds Cpu ONNX Runtime.