Create an inference model

Create an inference model from the cluster management console.

Before you begin

Before creating an inference model, confirm that the model has the required inference file.

  • TensorFlow inference requires that the model includes an inference.py file. Ensure that your model includes the required TensorFlow file, see Edit TensorFlow model for inference.

  • Caffe inference requires that the model includes an inference.prototxt file. Ensure that the model includes the required Caffe file, see Edit Caffe model for inference.

Procedure

  1. From the cluster management console, select Workload > Spark > Deep Learning.
  2. Click on the Models tab.
  3. Click on the training model that you want to create an inference model from.
  4. Select the Training tab.
  5. Click the training run that you want to create an inference model from and click Create Inference Model.
  6. Click Create to confirm model creation and create inference model.