Add a simple SmartSim experiment to use as a smoke test.
This experiment will launch an orchestrator. It will then attempt to store a tensor on the orchestrator at some key, and then retrieve the tensor stored at the same key.
If any of the ML backends (TF, Torch, ONNX) are installed, then for each backend, the test script will also create a simple ML model, store it on the orchestrator, and run the model with dummy a input tensor, and retrieve the output tensor.
Add a simple SmartSim experiment to use as a smoke test.
This experiment will launch an orchestrator. It will then attempt to store a tensor on the orchestrator at some key, and then retrieve the tensor stored at the same key.
If any of the ML backends (TF, Torch, ONNX) are installed, then for each backend, the test script will also create a simple ML model, store it on the orchestrator, and run the model with dummy a input tensor, and retrieve the output tensor.