consts

Documentation

Index

• Variables

Variables

var (
	CortexVersion      = "0.24.1" // CORTEX_VERSION
	CortexVersionMinor = "0.24"   // CORTEX_VERSION_MINOR

	SingleModelName = "_cortex_default"

	DefaultImagePythonPredictorCPU   = defaultDockerImage("python-predictor-cpu")
	DefaultImagePythonPredictorGPU   = defaultDockerImage("python-predictor-gpu")
	DefaultImagePythonPredictorInf   = defaultDockerImage("python-predictor-inf")
	DefaultImageTensorFlowServingCPU = defaultDockerImage("tensorflow-serving-cpu")
	DefaultImageTensorFlowServingGPU = defaultDockerImage("tensorflow-serving-gpu")
	DefaultImageTensorFlowServingInf = defaultDockerImage("tensorflow-serving-inf")
	DefaultImageTensorFlowPredictor  = defaultDockerImage("tensorflow-predictor")
	DefaultImageONNXPredictorCPU     = defaultDockerImage("onnx-predictor-cpu")
	DefaultImageONNXPredictorGPU     = defaultDockerImage("onnx-predictor-gpu")
	DefaultImagePathsSet             = strset.New(
		DefaultImagePythonPredictorCPU,
		DefaultImagePythonPredictorGPU,
		DefaultImagePythonPredictorInf,
		DefaultImageTensorFlowServingCPU,
		DefaultImageTensorFlowServingGPU,
		DefaultImageTensorFlowServingInf,
		DefaultImageTensorFlowPredictor,
		DefaultImageONNXPredictorCPU,
		DefaultImageONNXPredictorGPU,
	)

	MaxClassesPerMonitoringRequest = 20 // cloudwatch.GeMetricData can get up to 100 metrics per request, avoid multiple requests and have room for other stats
	DashboardTitle                 = "# cortex monitoring dashboard"
	DefaultMaxReplicaConcurrency   = int64(1024)
	NeuronCoresPerInf              = int64(4)
)