Triton Inference Server Unload a model

Request that a model be unloaded from Triton. Once unloaded the model will no longer be available for inference. This is a Triton extension to the KServe protocol.