Inference of SunoAI's bark model in pure C/C++ using ggml.

The main goal of bark.cpp is to synthesize audio from a textual input with the Bark model using 4-bit quantization on a MacBook.

Bark has essentially 4 components:

• Semantic model to encode the text input
• Coarse model
• Fine model
• Encoder (quantizer + decoder) to generate the waveform from the tokens

• Quantization
• FP16
• Swift package for iOS devices