Predict

Runs prediction (inference) on a model. Accepts input tensors and returns output tensors from the model's serving signature.