Backends
Qwodel wraps three quantization backends. Each backend page covers its supported formats, initialization parameters, and runtime overrides.
In this section
| Page | Backend | Hardware |
|---|---|---|
| GGUF | llama.cpp-compatible quantization | CPU |
| AWQ | Activation-Aware Weight Quantization | NVIDIA GPU |
| CoreML | Apple on-device inference | Apple Silicon / Intel Mac |
Not sure which to pick? See Choosing a Backend →