Documentation
¶
Index ¶
- Constants
- func BuildModelImageWithMemory(baseImageName, memoryLocation, newImageTag string) (result io.ReadSeeker, err error)
- func GetDockerClient() *client.Client
- func InferModuleProperties(sourcePath string) (id, name, description, schemaIn, schemaOut, configSpace string, err error)
- func LoadImage(imageFilePath string) (string, error)
- func RunContainerAndCollectOutput(imageName string, entrypoint, command []string, gpuDevices []string) (io.ReadCloser, error)
- func ValidateModel(modelImageName string, schemaStringIn, schemaStringOut, configSpace string, ...) (err error)
- type ConfigChoice
- type ConfigConst
- type ConfigElem
- type ConfigFloat
- type ConfigInt
- type ConfigMap
Constants ¶
const MntPrefix = "^^^"
MntPrefix must be placed before all command line arguments if they represent a local directory or file which we want to mount to the image.
Variables ¶
This section is empty.
Functions ¶
func BuildModelImageWithMemory ¶
func BuildModelImageWithMemory(baseImageName, memoryLocation, newImageTag string) (result io.ReadSeeker, err error)
BuildModelImageWithMemory takes a model image, copies the memory content to it and builds a new image from that.
func GetDockerClient ¶
GetDockerClient returns an instance of the Docker client.
func InferModuleProperties ¶
func InferModuleProperties(sourcePath string) (id, name, description, schemaIn, schemaOut, configSpace string, err error)
InferModuleProperties takes a module available on the local docker instance and tries to infer its basic properties such as id, name and description.
func RunContainerAndCollectOutput ¶
func RunContainerAndCollectOutput(imageName string, entrypoint, command []string, gpuDevices []string) (io.ReadCloser, error)
RunContainerAndCollectOutput runs a given image name and returns the standard output reader.
func ValidateModel ¶
func ValidateModel(modelImageName string, schemaStringIn, schemaStringOut, configSpace string, cleanup bool) (err error)
ValidateModel takes a model image and the input and output schema, generates a random input data set, runs train and predict on the model and verifies that the output data matches the output schema.
Types ¶
type ConfigElem ¶
type ConfigElem interface {
Sample() ConfigElem
Expand(rangeCount int) []ConfigElem
Dump() interface{}
}
ConfigElem is.
type ConfigFloat ¶
ConfigFloat is.
Source Files