Model Deployment

The model deploy file is used to leverage a trained model to perform inference on unknown set of node deviations.

class dlmfg.core.model_deployment.DeployModel[source]

The Deploy Model class is used to import a trained model and use it to infer on unknown data

get_model(model_path)[source]

get_model method is is used to retrieve the trained model from a given path

Parameters

model_path (str (required)) – Path to the trained model, ideally it should be same as the train model path output

model_inference(inference_data, inference_model, deploy_path, print_result=0, plot_result=0, get_cam_data=0, append_result=0)[source]

model_inference method is used to infer from unknown sample(s) using the trained model

Parameters
  • inference_data (numpy.array [samples*voxel_dim*voxel_dim*voxel_dim*deviation_channels] (required) (required)) – Unknown dataset having same structure as the train dataset

  • inference_model (keras.model (required)) – Trained model

  • print_result (int) – Flag to indicate if the result needs to be printed, 0 by default, change to 1 in case the results need to be printed on the console