starling.inference.generation.ensemble_encoder_backend

ensemble_encoder_backend(ensemble, device, batch_size, output_directory=None, model_manager=<starling.inference.model_loading.ModelManager object>, encoder_path=None, ddpm_path=None)