types

package
v0.3.2 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Feb 7, 2020 License: Apache-2.0, Apache-2.0 Imports: 9 Imported by: 0

Documentation

Index

Constants

This section is empty.

Variables

View Source
var (
	ErrNotFoundJobs = errors.New(`not found jobs under the assigned conditions.`)
	ErrTooManyJobs  = errors.New(`found jobs more than one,please use --version or --type to filter.`)
)
View Source
var KubeConfig string
View Source
var SERVING_CHARTS = map[string]string{
	"tensorflow-serving-0.2.0":        "Tensorflow",
	"tensorrt-inference-server-0.0.1": "TensorRT",
}

Functions

func DisplayGPUShareDetails added in v0.3.1

func DisplayGPUShareDetails(nodeInfos []*GPUShareNodeInfo)

func DisplayGPUShareSummary added in v0.3.1

func DisplayGPUShareSummary(nodeInfos []*GPUShareNodeInfo)

func GetGPUMemoryInPod added in v0.3.1

func GetGPUMemoryInPod(pod v1.Pod) int

Types

type DeviceInfo added in v0.3.1

type DeviceInfo struct {
	Pods        []v1.Pod
	UsedGPUMem  int
	TotalGPUMem int
	// contains filtered or unexported fields
}

func (*DeviceInfo) String added in v0.3.1

func (d *DeviceInfo) String() string

type GPUShareNodeInfo added in v0.3.1

type GPUShareNodeInfo struct {
	Devs     map[int]*DeviceInfo
	GpuCount int
	// contains filtered or unexported fields
}

func BuildAllGPUShareNodeInfos added in v0.3.1

func BuildAllGPUShareNodeInfos(allPods []v1.Pod, nodes []v1.Node) ([]*GPUShareNodeInfo, error)

For all GPUShare nodes,decide whether the memory of GPU is measured by MiB or GiB

func BuildGPUShareNodeInfo added in v0.3.1

func BuildGPUShareNodeInfo(allPods []v1.Pod, node v1.Node) (*GPUShareNodeInfo, error)

For one GPUShare node,decide whether the memory of GPU is measured by MiB or GiB

type Instance

type Instance struct {
	// the status of of instance
	Status string `json:"status"`
	// the name of instance
	Name string `json:"name"`
	// the age of instance
	Age string `json:"age"`
	// the node instance runs on
	Node string `json:"node"`
	// the instance is chief or not
	IsChief bool `json:"chief" yaml:"chief"`
}

type JobInfo

type JobInfo struct {
	// The name of the training job
	Name string `json:"name"`
	// The namespace of the training job
	Namespace string `json:"namespace"`
	// The time of the training job
	Duration string `json:"duration"`
	// The status of the training Job
	Status JobStatus `json:"status"`

	// The training type of the training job
	Trainer string `json:"trainer"`
	// The tensorboard of the training job
	Tensorboard string `json:"tensorboard,omitempty"`

	// The name of the chief Instance
	ChiefName string `json:"chiefName" yaml:"chiefName"`

	// The instances under the training job
	Instances []Instance `json:"instances"`

	// The priority of the training job
	Priority string `json:"priority"`
}

type JobStatus

type JobStatus string

all the kinds of JobStatus

const (
	// JobPending means the job is pending
	JobPending JobStatus = "PENDING"
	// JobRunning means the job is running
	JobRunning JobStatus = "RUNNING"
	// JobSucceeded means the job is Succeeded
	JobSucceeded JobStatus = "SUCCEEDED"
	// JobFailed means the job is failed
	JobFailed JobStatus = "FAILED"
)

type ServingType added in v0.3.0

type ServingType string
const (
	// tensorflow
	ServingTF ServingType = "TENSORFLOW"
	// tensorrt
	ServingTRT ServingType = "TENSORRT"
	// custom
	ServingCustom ServingType = "CUSTOM"
)

three serving types.

type TrainingJobInfo

type TrainingJobInfo struct {
	Name      string
	Type      string
	Namespace string
}

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL