def test_model_infer(inference_service_stub, model_infer_request): prediction = inference_service_stub.ModelInfer(model_infer_request) expected = pb.InferTensorContents(fp32_contents=[21.0]) assert len(prediction.outputs) == 1 assert prediction.outputs[0].contents == expected
async def test_model_infer(inference_service_stub, model_infer_request, model_name, model_version): model_infer_request.model_name = model_name if model_version is not None: model_infer_request.model_version = model_version else: model_infer_request.ClearField("model_version") prediction = await inference_service_stub.ModelInfer(model_infer_request) expected = pb.InferTensorContents(int64_contents=[6]) assert len(prediction.outputs) == 1 assert prediction.outputs[0].contents == expected
def test_modelinferresponse_from_types(inference_response): model_infer_response = ModelInferResponseConverter.from_types(inference_response) expected = pb.ModelInferResponse( model_name="sum-model", id="123", outputs=[ pb.ModelInferResponse.InferOutputTensor( name="output-0", datatype="FP32", shape=[1], contents=pb.InferTensorContents(fp32_contents=[21.0]), ) ], ) assert type(model_infer_response) is pb.ModelInferResponse assert json_format.MessageToDict(model_infer_response) == json_format.MessageToDict( expected )