wavenet-convnet

What problem does it solve?

WaveNet-based architectures enable real-time, high-fidelity neural audio synthesis and amp modeling by learning long-range temporal dependencies without recurrence, reducing latency while increasing realism.

Core Features & Use Cases

• Deterministic, real-time audio generation using dilated causal convolutions and residual gated blocks to simulate amp behavior and effects.
• Extensible architecture supports multiple layers, receptive field tuning, and real-time inference with caching for low latency.
• Use Case: Create an authentic guitar amp model or dynamic vocal synthesizer with responsive scheduling and minimal artifacts.

Quick Start

Train or deploy a WaveNet-convnet model with defined layers, dilation schedule, and causal padding to generate audio samples in real-time.