At runtime, the nn-inference-template vst3, component and standalone attempt to load the neural models from the user's application data directory. Therefore, you must copy the models as follows:
| source | path/to/release/extras/nn-inference-template/* |
|---|---|
| target | ~/Library/Application Support/nn-inference-template/* |
The macOS versions of the nn-inference-template are all bundled with the dynamic libraries for anira and the inference engines.