published on Wednesday, Nov 18, 2026 by Pulumi
published on Wednesday, Nov 18, 2026 by Pulumi
The TrainingPipeline orchestrates tasks associated with training a Model.
To get more information about TrainingPipeline, see:
- API documentation
- How-to Guides
Example Usage
Vertex Ai Training Pipeline Basic
import * as pulumi from "@pulumi/pulumi";
import * as gcp from "@pulumi/gcp";
const basicPipeline = new gcp.vertex.AiTrainingPipeline("basic_pipeline", {
displayName: "test-training-pipeline",
modelId: "test-model-id",
region: "us-central1",
trainingTaskDefinition: "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
trainingTaskInputs: JSON.stringify({
workerPoolSpecs: [{
machineSpec: {
machineType: "n1-standard-4",
},
replicaCount: "1",
diskSpec: {
bootDiskType: "pd-ssd",
bootDiskSizeGb: 100,
},
containerSpec: {
imageUri: "gcr.io/cloud-marketplace/google/debian11",
},
}],
}),
labels: {
env: "test",
},
});
import pulumi
import json
import pulumi_gcp as gcp
basic_pipeline = gcp.vertex.AiTrainingPipeline("basic_pipeline",
display_name="test-training-pipeline",
model_id="test-model-id",
region="us-central1",
training_task_definition="gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
training_task_inputs=json.dumps({
"workerPoolSpecs": [{
"machineSpec": {
"machineType": "n1-standard-4",
},
"replicaCount": "1",
"diskSpec": {
"bootDiskType": "pd-ssd",
"bootDiskSizeGb": 100,
},
"containerSpec": {
"imageUri": "gcr.io/cloud-marketplace/google/debian11",
},
}],
}),
labels={
"env": "test",
})
package main
import (
"encoding/json"
"github.com/pulumi/pulumi-gcp/sdk/v10/go/gcp/vertex"
"github.com/pulumi/pulumi/sdk/v3/go/pulumi"
)
func main() {
pulumi.Run(func(ctx *pulumi.Context) error {
tmpJSON0, err := json.Marshal(map[string][]map[string]interface{}{
"workerPoolSpecs": []map[string]interface{}{
map[string]interface{}{
"machineSpec": map[string]string{
"machineType": "n1-standard-4",
},
"replicaCount": "1",
"diskSpec": map[string]interface{}{
"bootDiskType": "pd-ssd",
"bootDiskSizeGb": 100,
},
"containerSpec": map[string]string{
"imageUri": "gcr.io/cloud-marketplace/google/debian11",
},
},
},
})
if err != nil {
return err
}
json0 := string(tmpJSON0)
_, err = vertex.NewAiTrainingPipeline(ctx, "basic_pipeline", &vertex.AiTrainingPipelineArgs{
DisplayName: pulumi.String("test-training-pipeline"),
ModelId: pulumi.String("test-model-id"),
Region: pulumi.String("us-central1"),
TrainingTaskDefinition: pulumi.String("gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml"),
TrainingTaskInputs: pulumi.String(json0),
Labels: pulumi.StringMap{
"env": pulumi.String("test"),
},
})
if err != nil {
return err
}
return nil
})
}
using System.Collections.Generic;
using System.Linq;
using System.Text.Json;
using Pulumi;
using Gcp = Pulumi.Gcp;
return await Deployment.RunAsync(() =>
{
var basicPipeline = new Gcp.Vertex.AiTrainingPipeline("basic_pipeline", new()
{
DisplayName = "test-training-pipeline",
ModelId = "test-model-id",
Region = "us-central1",
TrainingTaskDefinition = "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
TrainingTaskInputs = JsonSerializer.Serialize(new Dictionary<string, object?>
{
["workerPoolSpecs"] = new[]
{
new Dictionary<string, object?>
{
["machineSpec"] = new Dictionary<string, object?>
{
["machineType"] = "n1-standard-4",
},
["replicaCount"] = "1",
["diskSpec"] = new Dictionary<string, object?>
{
["bootDiskType"] = "pd-ssd",
["bootDiskSizeGb"] = 100,
},
["containerSpec"] = new Dictionary<string, object?>
{
["imageUri"] = "gcr.io/cloud-marketplace/google/debian11",
},
},
},
}),
Labels =
{
{ "env", "test" },
},
});
});
package generated_program;
import com.pulumi.Context;
import com.pulumi.Pulumi;
import com.pulumi.core.Output;
import com.pulumi.gcp.vertex.AiTrainingPipeline;
import com.pulumi.gcp.vertex.AiTrainingPipelineArgs;
import static com.pulumi.codegen.internal.Serialization.*;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Map;
import java.io.File;
import java.nio.file.Files;
import java.nio.file.Paths;
public class App {
public static void main(String[] args) {
Pulumi.run(App::stack);
}
public static void stack(Context ctx) {
var basicPipeline = new AiTrainingPipeline("basicPipeline", AiTrainingPipelineArgs.builder()
.displayName("test-training-pipeline")
.modelId("test-model-id")
.region("us-central1")
.trainingTaskDefinition("gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml")
.trainingTaskInputs(serializeJson(
jsonObject(
jsonProperty("workerPoolSpecs", jsonArray(jsonObject(
jsonProperty("machineSpec", jsonObject(
jsonProperty("machineType", "n1-standard-4")
)),
jsonProperty("replicaCount", "1"),
jsonProperty("diskSpec", jsonObject(
jsonProperty("bootDiskType", "pd-ssd"),
jsonProperty("bootDiskSizeGb", 100)
)),
jsonProperty("containerSpec", jsonObject(
jsonProperty("imageUri", "gcr.io/cloud-marketplace/google/debian11")
))
)))
)))
.labels(Map.of("env", "test"))
.build());
}
}
resources:
basicPipeline:
type: gcp:vertex:AiTrainingPipeline
name: basic_pipeline
properties:
displayName: test-training-pipeline
modelId: test-model-id
region: us-central1
trainingTaskDefinition: gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml
trainingTaskInputs:
fn::toJSON:
workerPoolSpecs:
- machineSpec:
machineType: n1-standard-4
replicaCount: '1'
diskSpec:
bootDiskType: pd-ssd
bootDiskSizeGb: 100
containerSpec:
imageUri: gcr.io/cloud-marketplace/google/debian11
labels:
env: test
terraform {
required_providers {
gcp = {
source = "pulumi/gcp"
}
}
}
resource "gcp_vertex_ai_training_pipeline" "basic_pipeline" {
lifecycle {
create_before_destroy = true
}
display_name = "test-training-pipeline"
model_id = "test-model-id"
region = "us-central1"
training_task_definition = "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml"
training_task_inputs = jsonencode({
"workerPoolSpecs" = [{
"machineSpec" = {
"machineType" = "n1-standard-4"
}
"replicaCount" = "1"
"diskSpec" = {
"bootDiskType" = "pd-ssd"
"bootDiskSizeGb" = 100
}
"containerSpec" = {
"imageUri" = "gcr.io/cloud-marketplace/google/debian11"
}
}]
})
labels = {
"env" = "test"
}
}
Vertex Ai Training Pipeline Model
import * as pulumi from "@pulumi/pulumi";
import * as gcp from "@pulumi/gcp";
const bucket = new gcp.storage.Bucket("bucket", {
name: "test-model-bucket",
location: "us-central1",
forceDestroy: true,
uniformBucketLevelAccess: true,
});
const modelPipeline = new gcp.vertex.AiTrainingPipeline("model_pipeline", {
displayName: "test-training-pipeline-model",
region: "us-central1",
trainingTaskDefinition: "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
trainingTaskInputs: JSON.stringify({
workerPoolSpecs: [{
machineSpec: {
machineType: "n1-standard-4",
},
replicaCount: "1",
diskSpec: {
bootDiskType: "pd-ssd",
bootDiskSizeGb: 100,
},
containerSpec: {
imageUri: "gcr.io/cloud-marketplace/google/debian11",
},
}],
}),
modelToUpload: {
displayName: "test-model",
description: "A custom model uploaded from training pipeline",
versionDescription: "v1 initial release",
versionAliases: [
"v1",
"latest",
],
artifactUri: pulumi.interpolate`gs://${bucket.name}/model`,
containerSpec: {
imageUri: "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest",
commands: ["python"],
args: [
"-m",
"server",
],
predictRoute: "/predict",
healthRoute: "/health",
envs: [{
name: "APP_ENV",
value: "production",
}],
ports: [{
containerPort: 8080,
}],
grpcPorts: [{
containerPort: 8081,
}],
deploymentTimeout: "1800s",
sharedMemorySizeMb: "128",
startupProbe: {
exec: {
commands: [
"cat",
"/tmp/ready",
],
},
periodSeconds: 10,
timeoutSeconds: 10,
},
healthProbe: {
exec: {
commands: [
"cat",
"/tmp/healthy",
],
},
periodSeconds: 10,
timeoutSeconds: 10,
},
},
labels: {
model_type: "custom",
},
explanationSpec: {
parameters: JSON.stringify({
sampledShapleyAttribution: {
pathCount: 5,
},
}),
metadata: JSON.stringify({
inputs: {
features: {},
},
outputs: {
predictions: {},
},
}),
},
},
modelId: "test-model-id",
parentModel: "test-parent-model",
labels: {
env: "test",
},
});
import pulumi
import json
import pulumi_gcp as gcp
bucket = gcp.storage.Bucket("bucket",
name="test-model-bucket",
location="us-central1",
force_destroy=True,
uniform_bucket_level_access=True)
model_pipeline = gcp.vertex.AiTrainingPipeline("model_pipeline",
display_name="test-training-pipeline-model",
region="us-central1",
training_task_definition="gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
training_task_inputs=json.dumps({
"workerPoolSpecs": [{
"machineSpec": {
"machineType": "n1-standard-4",
},
"replicaCount": "1",
"diskSpec": {
"bootDiskType": "pd-ssd",
"bootDiskSizeGb": 100,
},
"containerSpec": {
"imageUri": "gcr.io/cloud-marketplace/google/debian11",
},
}],
}),
model_to_upload={
"display_name": "test-model",
"description": "A custom model uploaded from training pipeline",
"version_description": "v1 initial release",
"version_aliases": [
"v1",
"latest",
],
"artifact_uri": bucket.name.apply(lambda name: f"gs://{name}/model"),
"container_spec": {
"image_uri": "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest",
"commands": ["python"],
"args": [
"-m",
"server",
],
"predict_route": "/predict",
"health_route": "/health",
"envs": [{
"name": "APP_ENV",
"value": "production",
}],
"ports": [{
"container_port": 8080,
}],
"grpc_ports": [{
"container_port": 8081,
}],
"deployment_timeout": "1800s",
"shared_memory_size_mb": "128",
"startup_probe": {
"exec_": {
"commands": [
"cat",
"/tmp/ready",
],
},
"period_seconds": 10,
"timeout_seconds": 10,
},
"health_probe": {
"exec_": {
"commands": [
"cat",
"/tmp/healthy",
],
},
"period_seconds": 10,
"timeout_seconds": 10,
},
},
"labels": {
"model_type": "custom",
},
"explanation_spec": {
"parameters": json.dumps({
"sampledShapleyAttribution": {
"pathCount": 5,
},
}),
"metadata": json.dumps({
"inputs": {
"features": {},
},
"outputs": {
"predictions": {},
},
}),
},
},
model_id="test-model-id",
parent_model="test-parent-model",
labels={
"env": "test",
})
package main
import (
"encoding/json"
"fmt"
"github.com/pulumi/pulumi-gcp/sdk/v10/go/gcp/storage"
"github.com/pulumi/pulumi-gcp/sdk/v10/go/gcp/vertex"
"github.com/pulumi/pulumi/sdk/v3/go/pulumi"
)
func main() {
pulumi.Run(func(ctx *pulumi.Context) error {
bucket, err := storage.NewBucket(ctx, "bucket", &storage.BucketArgs{
Name: pulumi.String("test-model-bucket"),
Location: pulumi.String("us-central1"),
ForceDestroy: pulumi.Bool(true),
UniformBucketLevelAccess: pulumi.Bool(true),
})
if err != nil {
return err
}
tmpJSON0, err := json.Marshal(map[string][]map[string]interface{}{
"workerPoolSpecs": []map[string]interface{}{
map[string]interface{}{
"machineSpec": map[string]string{
"machineType": "n1-standard-4",
},
"replicaCount": "1",
"diskSpec": map[string]interface{}{
"bootDiskType": "pd-ssd",
"bootDiskSizeGb": 100,
},
"containerSpec": map[string]string{
"imageUri": "gcr.io/cloud-marketplace/google/debian11",
},
},
},
})
if err != nil {
return err
}
json0 := string(tmpJSON0)
tmpJSON1, err := json.Marshal(map[string]map[string]int{
"sampledShapleyAttribution": map[string]int{
"pathCount": 5,
},
})
if err != nil {
return err
}
json1 := string(tmpJSON1)
tmpJSON2, err := json.Marshal(map[string]map[string]map[string]interface{}{
"inputs": map[string]map[string]interface{}{
"features": map[string]interface{}{},
},
"outputs": map[string]map[string]interface{}{
"predictions": map[string]interface{}{},
},
})
if err != nil {
return err
}
json2 := string(tmpJSON2)
_, err = vertex.NewAiTrainingPipeline(ctx, "model_pipeline", &vertex.AiTrainingPipelineArgs{
DisplayName: pulumi.String("test-training-pipeline-model"),
Region: pulumi.String("us-central1"),
TrainingTaskDefinition: pulumi.String("gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml"),
TrainingTaskInputs: pulumi.String(json0),
ModelToUpload: &vertex.AiTrainingPipelineModelToUploadArgs{
DisplayName: pulumi.String("test-model"),
Description: pulumi.String("A custom model uploaded from training pipeline"),
VersionDescription: pulumi.String("v1 initial release"),
VersionAliases: pulumi.StringArray{
pulumi.String("v1"),
pulumi.String("latest"),
},
ArtifactUri: bucket.Name.ApplyT(func(name string) (string, error) {
return fmt.Sprintf("gs://%v/model", name), nil
}).(pulumi.StringOutput),
ContainerSpec: &vertex.AiTrainingPipelineModelToUploadContainerSpecArgs{
ImageUri: pulumi.String("us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest"),
Commands: pulumi.StringArray{
pulumi.String("python"),
},
Args: pulumi.StringArray{
pulumi.String("-m"),
pulumi.String("server"),
},
PredictRoute: pulumi.String("/predict"),
HealthRoute: pulumi.String("/health"),
Envs: vertex.AiTrainingPipelineModelToUploadContainerSpecEnvArray{
&vertex.AiTrainingPipelineModelToUploadContainerSpecEnvArgs{
Name: pulumi.String("APP_ENV"),
Value: pulumi.String("production"),
},
},
Ports: vertex.AiTrainingPipelineModelToUploadContainerSpecPortArray{
&vertex.AiTrainingPipelineModelToUploadContainerSpecPortArgs{
ContainerPort: pulumi.Int(8080),
},
},
GrpcPorts: vertex.AiTrainingPipelineModelToUploadContainerSpecGrpcPortArray{
&vertex.AiTrainingPipelineModelToUploadContainerSpecGrpcPortArgs{
ContainerPort: pulumi.Int(8081),
},
},
DeploymentTimeout: pulumi.String("1800s"),
SharedMemorySizeMb: pulumi.String("128"),
StartupProbe: &vertex.AiTrainingPipelineModelToUploadContainerSpecStartupProbeArgs{
Exec: &vertex.AiTrainingPipelineModelToUploadContainerSpecStartupProbeExecArgs{
Commands: pulumi.StringArray{
pulumi.String("cat"),
pulumi.String("/tmp/ready"),
},
},
PeriodSeconds: pulumi.Int(10),
TimeoutSeconds: pulumi.Int(10),
},
HealthProbe: &vertex.AiTrainingPipelineModelToUploadContainerSpecHealthProbeArgs{
Exec: &vertex.AiTrainingPipelineModelToUploadContainerSpecHealthProbeExecArgs{
Commands: pulumi.StringArray{
pulumi.String("cat"),
pulumi.String("/tmp/healthy"),
},
},
PeriodSeconds: pulumi.Int(10),
TimeoutSeconds: pulumi.Int(10),
},
},
Labels: pulumi.StringMap{
"model_type": pulumi.String("custom"),
},
ExplanationSpec: &vertex.AiTrainingPipelineModelToUploadExplanationSpecArgs{
Parameters: pulumi.String(json1),
Metadata: pulumi.String(json2),
},
},
ModelId: pulumi.String("test-model-id"),
ParentModel: pulumi.String("test-parent-model"),
Labels: pulumi.StringMap{
"env": pulumi.String("test"),
},
})
if err != nil {
return err
}
return nil
})
}
using System.Collections.Generic;
using System.Linq;
using System.Text.Json;
using Pulumi;
using Gcp = Pulumi.Gcp;
return await Deployment.RunAsync(() =>
{
var bucket = new Gcp.Storage.Bucket("bucket", new()
{
Name = "test-model-bucket",
Location = "us-central1",
ForceDestroy = true,
UniformBucketLevelAccess = true,
});
var modelPipeline = new Gcp.Vertex.AiTrainingPipeline("model_pipeline", new()
{
DisplayName = "test-training-pipeline-model",
Region = "us-central1",
TrainingTaskDefinition = "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
TrainingTaskInputs = JsonSerializer.Serialize(new Dictionary<string, object?>
{
["workerPoolSpecs"] = new[]
{
new Dictionary<string, object?>
{
["machineSpec"] = new Dictionary<string, object?>
{
["machineType"] = "n1-standard-4",
},
["replicaCount"] = "1",
["diskSpec"] = new Dictionary<string, object?>
{
["bootDiskType"] = "pd-ssd",
["bootDiskSizeGb"] = 100,
},
["containerSpec"] = new Dictionary<string, object?>
{
["imageUri"] = "gcr.io/cloud-marketplace/google/debian11",
},
},
},
}),
ModelToUpload = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadArgs
{
DisplayName = "test-model",
Description = "A custom model uploaded from training pipeline",
VersionDescription = "v1 initial release",
VersionAliases = new[]
{
"v1",
"latest",
},
ArtifactUri = bucket.Name.Apply(name => $"gs://{name}/model"),
ContainerSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecArgs
{
ImageUri = "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest",
Commands = new[]
{
"python",
},
Args = new[]
{
"-m",
"server",
},
PredictRoute = "/predict",
HealthRoute = "/health",
Envs = new[]
{
new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecEnvArgs
{
Name = "APP_ENV",
Value = "production",
},
},
Ports = new[]
{
new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecPortArgs
{
ContainerPort = 8080,
},
},
GrpcPorts = new[]
{
new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecGrpcPortArgs
{
ContainerPort = 8081,
},
},
DeploymentTimeout = "1800s",
SharedMemorySizeMb = "128",
StartupProbe = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecStartupProbeArgs
{
Exec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecStartupProbeExecArgs
{
Commands = new[]
{
"cat",
"/tmp/ready",
},
},
PeriodSeconds = 10,
TimeoutSeconds = 10,
},
HealthProbe = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecHealthProbeArgs
{
Exec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecHealthProbeExecArgs
{
Commands = new[]
{
"cat",
"/tmp/healthy",
},
},
PeriodSeconds = 10,
TimeoutSeconds = 10,
},
},
Labels =
{
{ "model_type", "custom" },
},
ExplanationSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadExplanationSpecArgs
{
Parameters = JsonSerializer.Serialize(new Dictionary<string, object?>
{
["sampledShapleyAttribution"] = new Dictionary<string, object?>
{
["pathCount"] = 5,
},
}),
Metadata = JsonSerializer.Serialize(new Dictionary<string, object?>
{
["inputs"] = new Dictionary<string, object?>
{
["features"] = new Dictionary<string, object?>
{
},
},
["outputs"] = new Dictionary<string, object?>
{
["predictions"] = new Dictionary<string, object?>
{
},
},
}),
},
},
ModelId = "test-model-id",
ParentModel = "test-parent-model",
Labels =
{
{ "env", "test" },
},
});
});
package generated_program;
import com.pulumi.Context;
import com.pulumi.Pulumi;
import com.pulumi.core.Output;
import com.pulumi.gcp.storage.Bucket;
import com.pulumi.gcp.storage.BucketArgs;
import com.pulumi.gcp.vertex.AiTrainingPipeline;
import com.pulumi.gcp.vertex.AiTrainingPipelineArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecEnvArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecPortArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecGrpcPortArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecStartupProbeArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecStartupProbeExecArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecHealthProbeArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecHealthProbeExecArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadExplanationSpecArgs;
import static com.pulumi.codegen.internal.Serialization.*;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Map;
import java.io.File;
import java.nio.file.Files;
import java.nio.file.Paths;
public class App {
public static void main(String[] args) {
Pulumi.run(App::stack);
}
public static void stack(Context ctx) {
var bucket = new Bucket("bucket", BucketArgs.builder()
.name("test-model-bucket")
.location("us-central1")
.forceDestroy(true)
.uniformBucketLevelAccess(true)
.build());
var modelPipeline = new AiTrainingPipeline("modelPipeline", AiTrainingPipelineArgs.builder()
.displayName("test-training-pipeline-model")
.region("us-central1")
.trainingTaskDefinition("gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml")
.trainingTaskInputs(serializeJson(
jsonObject(
jsonProperty("workerPoolSpecs", jsonArray(jsonObject(
jsonProperty("machineSpec", jsonObject(
jsonProperty("machineType", "n1-standard-4")
)),
jsonProperty("replicaCount", "1"),
jsonProperty("diskSpec", jsonObject(
jsonProperty("bootDiskType", "pd-ssd"),
jsonProperty("bootDiskSizeGb", 100)
)),
jsonProperty("containerSpec", jsonObject(
jsonProperty("imageUri", "gcr.io/cloud-marketplace/google/debian11")
))
)))
)))
.modelToUpload(AiTrainingPipelineModelToUploadArgs.builder()
.displayName("test-model")
.description("A custom model uploaded from training pipeline")
.versionDescription("v1 initial release")
.versionAliases(
"v1",
"latest")
.artifactUri(bucket.name().applyValue(_name -> String.format("gs://%s/model", _name)))
.containerSpec(AiTrainingPipelineModelToUploadContainerSpecArgs.builder()
.imageUri("us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest")
.commands("python")
.args(
"-m",
"server")
.predictRoute("/predict")
.healthRoute("/health")
.envs(AiTrainingPipelineModelToUploadContainerSpecEnvArgs.builder()
.name("APP_ENV")
.value("production")
.build())
.ports(AiTrainingPipelineModelToUploadContainerSpecPortArgs.builder()
.containerPort(8080)
.build())
.grpcPorts(AiTrainingPipelineModelToUploadContainerSpecGrpcPortArgs.builder()
.containerPort(8081)
.build())
.deploymentTimeout("1800s")
.sharedMemorySizeMb("128")
.startupProbe(AiTrainingPipelineModelToUploadContainerSpecStartupProbeArgs.builder()
.exec(AiTrainingPipelineModelToUploadContainerSpecStartupProbeExecArgs.builder()
.commands(
"cat",
"/tmp/ready")
.build())
.periodSeconds(10)
.timeoutSeconds(10)
.build())
.healthProbe(AiTrainingPipelineModelToUploadContainerSpecHealthProbeArgs.builder()
.exec(AiTrainingPipelineModelToUploadContainerSpecHealthProbeExecArgs.builder()
.commands(
"cat",
"/tmp/healthy")
.build())
.periodSeconds(10)
.timeoutSeconds(10)
.build())
.build())
.labels(Map.of("model_type", "custom"))
.explanationSpec(AiTrainingPipelineModelToUploadExplanationSpecArgs.builder()
.parameters(serializeJson(
jsonObject(
jsonProperty("sampledShapleyAttribution", jsonObject(
jsonProperty("pathCount", 5)
))
)))
.metadata(serializeJson(
jsonObject(
jsonProperty("inputs", jsonObject(
jsonProperty("features", jsonObject(
))
)),
jsonProperty("outputs", jsonObject(
jsonProperty("predictions", jsonObject(
))
))
)))
.build())
.build())
.modelId("test-model-id")
.parentModel("test-parent-model")
.labels(Map.of("env", "test"))
.build());
}
}
resources:
modelPipeline:
type: gcp:vertex:AiTrainingPipeline
name: model_pipeline
properties:
displayName: test-training-pipeline-model
region: us-central1
trainingTaskDefinition: gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml
trainingTaskInputs:
fn::toJSON:
workerPoolSpecs:
- machineSpec:
machineType: n1-standard-4
replicaCount: '1'
diskSpec:
bootDiskType: pd-ssd
bootDiskSizeGb: 100
containerSpec:
imageUri: gcr.io/cloud-marketplace/google/debian11
modelToUpload:
displayName: test-model
description: A custom model uploaded from training pipeline
versionDescription: v1 initial release
versionAliases:
- v1
- latest
artifactUri: gs://${bucket.name}/model
containerSpec:
imageUri: us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest
commands:
- python
args:
- -m
- server
predictRoute: /predict
healthRoute: /health
envs:
- name: APP_ENV
value: production
ports:
- containerPort: 8080
grpcPorts:
- containerPort: 8081
deploymentTimeout: 1800s
sharedMemorySizeMb: '128'
startupProbe:
exec:
commands:
- cat
- /tmp/ready
periodSeconds: 10
timeoutSeconds: 10
healthProbe:
exec:
commands:
- cat
- /tmp/healthy
periodSeconds: 10
timeoutSeconds: 10
labels:
model_type: custom
explanationSpec:
parameters:
fn::toJSON:
sampledShapleyAttribution:
pathCount: 5
metadata:
fn::toJSON:
inputs:
features: {}
outputs:
predictions: {}
modelId: test-model-id
parentModel: test-parent-model
labels:
env: test
bucket:
type: gcp:storage:Bucket
properties:
name: test-model-bucket
location: us-central1
forceDestroy: true
uniformBucketLevelAccess: true
terraform {
required_providers {
gcp = {
source = "pulumi/gcp"
}
}
}
resource "gcp_vertex_ai_training_pipeline" "model_pipeline" {
lifecycle {
create_before_destroy = true
}
display_name = "test-training-pipeline-model"
region = "us-central1"
training_task_definition = "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml"
training_task_inputs = jsonencode({
"workerPoolSpecs" = [{
"machineSpec" = {
"machineType" = "n1-standard-4"
}
"replicaCount" = "1"
"diskSpec" = {
"bootDiskType" = "pd-ssd"
"bootDiskSizeGb" = 100
}
"containerSpec" = {
"imageUri" = "gcr.io/cloud-marketplace/google/debian11"
}
}]
})
model_to_upload = {
display_name = "test-model"
description = "A custom model uploaded from training pipeline"
version_description = "v1 initial release"
version_aliases = ["v1", "latest"]
artifact_uri ="gs://${gcp_storage_bucket.bucket.name}/model"
container_spec = {
image_uri = "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest"
commands = ["python"]
args = ["-m", "server"]
predict_route = "/predict"
health_route = "/health"
envs = [{
name = "APP_ENV"
value = "production"
}]
ports = [{
container_port = 8080
}]
grpc_ports = [{
container_port = 8081
}]
deployment_timeout = "1800s"
shared_memory_size_mb = "128"
startup_probe = {
exec = {
commands = ["cat", "/tmp/ready"]
}
period_seconds = 10
timeout_seconds = 10
}
health_probe = {
exec = {
commands = ["cat", "/tmp/healthy"]
}
period_seconds = 10
timeout_seconds = 10
}
}
labels = {
"model_type" = "custom"
}
explanation_spec = {
parameters = jsonencode({
"sampledShapleyAttribution" = {
"pathCount" = 5
}
})
metadata = jsonencode({
"inputs" = {
"features" = {}
}
"outputs" = {
"predictions" = {}
}
})
}
}
model_id = "test-model-id"
parent_model = "test-parent-model"
labels = {
"env" = "test"
}
}
resource "gcp_storage_bucket" "bucket" {
lifecycle {
create_before_destroy = true
}
name = "test-model-bucket"
location = "us-central1"
force_destroy = true
uniform_bucket_level_access = true
}
Vertex Ai Training Pipeline Dataset
import * as pulumi from "@pulumi/pulumi";
import * as gcp from "@pulumi/gcp";
const dataset = new gcp.vertex.AiDataset("dataset", {
displayName: "test-dataset",
metadataSchemaUri: "gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml",
region: "us-central1",
});
const bucket = new gcp.storage.Bucket("bucket", {
name: "test-bucket",
location: "us-central1",
forceDestroy: true,
uniformBucketLevelAccess: true,
});
const datasetPipeline = new gcp.vertex.AiTrainingPipeline("dataset_pipeline", {
displayName: "test-training-pipeline-dataset",
modelId: "test-model-id",
region: "us-central1",
trainingTaskDefinition: "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
trainingTaskInputs: JSON.stringify({
workerPoolSpecs: [{
machineSpec: {
machineType: "n1-standard-4",
},
replicaCount: "1",
diskSpec: {
bootDiskType: "pd-ssd",
bootDiskSizeGb: 100,
},
containerSpec: {
imageUri: "gcr.io/cloud-marketplace/google/debian11",
},
}],
}),
inputDataConfig: {
datasetId: dataset.name,
annotationSchemaUri: "gs://google-cloud-aiplatform/schema/dataset/annotation/image_classification_1.0.0.yaml",
annotationsFilter: "annotation_filter_key:annotation_filter_value",
persistMlUseAssignment: false,
filterSplit: {
trainingFilter: "labels.test:training",
validationFilter: "labels.test:validation",
testFilter: "labels.test:test",
},
gcsDestination: {
outputUriPrefix: pulumi.interpolate`gs://${bucket.name}/dataset-export`,
},
},
labels: {
env: "test",
},
});
import pulumi
import json
import pulumi_gcp as gcp
dataset = gcp.vertex.AiDataset("dataset",
display_name="test-dataset",
metadata_schema_uri="gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml",
region="us-central1")
bucket = gcp.storage.Bucket("bucket",
name="test-bucket",
location="us-central1",
force_destroy=True,
uniform_bucket_level_access=True)
dataset_pipeline = gcp.vertex.AiTrainingPipeline("dataset_pipeline",
display_name="test-training-pipeline-dataset",
model_id="test-model-id",
region="us-central1",
training_task_definition="gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
training_task_inputs=json.dumps({
"workerPoolSpecs": [{
"machineSpec": {
"machineType": "n1-standard-4",
},
"replicaCount": "1",
"diskSpec": {
"bootDiskType": "pd-ssd",
"bootDiskSizeGb": 100,
},
"containerSpec": {
"imageUri": "gcr.io/cloud-marketplace/google/debian11",
},
}],
}),
input_data_config={
"dataset_id": dataset.name,
"annotation_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/annotation/image_classification_1.0.0.yaml",
"annotations_filter": "annotation_filter_key:annotation_filter_value",
"persist_ml_use_assignment": False,
"filter_split": {
"training_filter": "labels.test:training",
"validation_filter": "labels.test:validation",
"test_filter": "labels.test:test",
},
"gcs_destination": {
"output_uri_prefix": bucket.name.apply(lambda name: f"gs://{name}/dataset-export"),
},
},
labels={
"env": "test",
})
package main
import (
"encoding/json"
"fmt"
"github.com/pulumi/pulumi-gcp/sdk/v10/go/gcp/storage"
"github.com/pulumi/pulumi-gcp/sdk/v10/go/gcp/vertex"
"github.com/pulumi/pulumi/sdk/v3/go/pulumi"
)
func main() {
pulumi.Run(func(ctx *pulumi.Context) error {
dataset, err := vertex.NewAiDataset(ctx, "dataset", &vertex.AiDatasetArgs{
DisplayName: pulumi.String("test-dataset"),
MetadataSchemaUri: pulumi.String("gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml"),
Region: pulumi.String("us-central1"),
})
if err != nil {
return err
}
bucket, err := storage.NewBucket(ctx, "bucket", &storage.BucketArgs{
Name: pulumi.String("test-bucket"),
Location: pulumi.String("us-central1"),
ForceDestroy: pulumi.Bool(true),
UniformBucketLevelAccess: pulumi.Bool(true),
})
if err != nil {
return err
}
tmpJSON0, err := json.Marshal(map[string][]map[string]interface{}{
"workerPoolSpecs": []map[string]interface{}{
map[string]interface{}{
"machineSpec": map[string]string{
"machineType": "n1-standard-4",
},
"replicaCount": "1",
"diskSpec": map[string]interface{}{
"bootDiskType": "pd-ssd",
"bootDiskSizeGb": 100,
},
"containerSpec": map[string]string{
"imageUri": "gcr.io/cloud-marketplace/google/debian11",
},
},
},
})
if err != nil {
return err
}
json0 := string(tmpJSON0)
_, err = vertex.NewAiTrainingPipeline(ctx, "dataset_pipeline", &vertex.AiTrainingPipelineArgs{
DisplayName: pulumi.String("test-training-pipeline-dataset"),
ModelId: pulumi.String("test-model-id"),
Region: pulumi.String("us-central1"),
TrainingTaskDefinition: pulumi.String("gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml"),
TrainingTaskInputs: pulumi.String(json0),
InputDataConfig: &vertex.AiTrainingPipelineInputDataConfigArgs{
DatasetId: dataset.Name,
AnnotationSchemaUri: pulumi.String("gs://google-cloud-aiplatform/schema/dataset/annotation/image_classification_1.0.0.yaml"),
AnnotationsFilter: pulumi.String("annotation_filter_key:annotation_filter_value"),
PersistMlUseAssignment: pulumi.Bool(false),
FilterSplit: &vertex.AiTrainingPipelineInputDataConfigFilterSplitArgs{
TrainingFilter: pulumi.String("labels.test:training"),
ValidationFilter: pulumi.String("labels.test:validation"),
TestFilter: pulumi.String("labels.test:test"),
},
GcsDestination: &vertex.AiTrainingPipelineInputDataConfigGcsDestinationArgs{
OutputUriPrefix: bucket.Name.ApplyT(func(name string) (string, error) {
return fmt.Sprintf("gs://%v/dataset-export", name), nil
}).(pulumi.StringOutput),
},
},
Labels: pulumi.StringMap{
"env": pulumi.String("test"),
},
})
if err != nil {
return err
}
return nil
})
}
using System.Collections.Generic;
using System.Linq;
using System.Text.Json;
using Pulumi;
using Gcp = Pulumi.Gcp;
return await Deployment.RunAsync(() =>
{
var dataset = new Gcp.Vertex.AiDataset("dataset", new()
{
DisplayName = "test-dataset",
MetadataSchemaUri = "gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml",
Region = "us-central1",
});
var bucket = new Gcp.Storage.Bucket("bucket", new()
{
Name = "test-bucket",
Location = "us-central1",
ForceDestroy = true,
UniformBucketLevelAccess = true,
});
var datasetPipeline = new Gcp.Vertex.AiTrainingPipeline("dataset_pipeline", new()
{
DisplayName = "test-training-pipeline-dataset",
ModelId = "test-model-id",
Region = "us-central1",
TrainingTaskDefinition = "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
TrainingTaskInputs = JsonSerializer.Serialize(new Dictionary<string, object?>
{
["workerPoolSpecs"] = new[]
{
new Dictionary<string, object?>
{
["machineSpec"] = new Dictionary<string, object?>
{
["machineType"] = "n1-standard-4",
},
["replicaCount"] = "1",
["diskSpec"] = new Dictionary<string, object?>
{
["bootDiskType"] = "pd-ssd",
["bootDiskSizeGb"] = 100,
},
["containerSpec"] = new Dictionary<string, object?>
{
["imageUri"] = "gcr.io/cloud-marketplace/google/debian11",
},
},
},
}),
InputDataConfig = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigArgs
{
DatasetId = dataset.Name,
AnnotationSchemaUri = "gs://google-cloud-aiplatform/schema/dataset/annotation/image_classification_1.0.0.yaml",
AnnotationsFilter = "annotation_filter_key:annotation_filter_value",
PersistMlUseAssignment = false,
FilterSplit = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigFilterSplitArgs
{
TrainingFilter = "labels.test:training",
ValidationFilter = "labels.test:validation",
TestFilter = "labels.test:test",
},
GcsDestination = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigGcsDestinationArgs
{
OutputUriPrefix = bucket.Name.Apply(name => $"gs://{name}/dataset-export"),
},
},
Labels =
{
{ "env", "test" },
},
});
});
package generated_program;
import com.pulumi.Context;
import com.pulumi.Pulumi;
import com.pulumi.core.Output;
import com.pulumi.gcp.vertex.AiDataset;
import com.pulumi.gcp.vertex.AiDatasetArgs;
import com.pulumi.gcp.storage.Bucket;
import com.pulumi.gcp.storage.BucketArgs;
import com.pulumi.gcp.vertex.AiTrainingPipeline;
import com.pulumi.gcp.vertex.AiTrainingPipelineArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineInputDataConfigArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineInputDataConfigFilterSplitArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineInputDataConfigGcsDestinationArgs;
import static com.pulumi.codegen.internal.Serialization.*;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Map;
import java.io.File;
import java.nio.file.Files;
import java.nio.file.Paths;
public class App {
public static void main(String[] args) {
Pulumi.run(App::stack);
}
public static void stack(Context ctx) {
var dataset = new AiDataset("dataset", AiDatasetArgs.builder()
.displayName("test-dataset")
.metadataSchemaUri("gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml")
.region("us-central1")
.build());
var bucket = new Bucket("bucket", BucketArgs.builder()
.name("test-bucket")
.location("us-central1")
.forceDestroy(true)
.uniformBucketLevelAccess(true)
.build());
var datasetPipeline = new AiTrainingPipeline("datasetPipeline", AiTrainingPipelineArgs.builder()
.displayName("test-training-pipeline-dataset")
.modelId("test-model-id")
.region("us-central1")
.trainingTaskDefinition("gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml")
.trainingTaskInputs(serializeJson(
jsonObject(
jsonProperty("workerPoolSpecs", jsonArray(jsonObject(
jsonProperty("machineSpec", jsonObject(
jsonProperty("machineType", "n1-standard-4")
)),
jsonProperty("replicaCount", "1"),
jsonProperty("diskSpec", jsonObject(
jsonProperty("bootDiskType", "pd-ssd"),
jsonProperty("bootDiskSizeGb", 100)
)),
jsonProperty("containerSpec", jsonObject(
jsonProperty("imageUri", "gcr.io/cloud-marketplace/google/debian11")
))
)))
)))
.inputDataConfig(AiTrainingPipelineInputDataConfigArgs.builder()
.datasetId(dataset.name())
.annotationSchemaUri("gs://google-cloud-aiplatform/schema/dataset/annotation/image_classification_1.0.0.yaml")
.annotationsFilter("annotation_filter_key:annotation_filter_value")
.persistMlUseAssignment(false)
.filterSplit(AiTrainingPipelineInputDataConfigFilterSplitArgs.builder()
.trainingFilter("labels.test:training")
.validationFilter("labels.test:validation")
.testFilter("labels.test:test")
.build())
.gcsDestination(AiTrainingPipelineInputDataConfigGcsDestinationArgs.builder()
.outputUriPrefix(bucket.name().applyValue(_name -> String.format("gs://%s/dataset-export", _name)))
.build())
.build())
.labels(Map.of("env", "test"))
.build());
}
}
resources:
datasetPipeline:
type: gcp:vertex:AiTrainingPipeline
name: dataset_pipeline
properties:
displayName: test-training-pipeline-dataset
modelId: test-model-id
region: us-central1
trainingTaskDefinition: gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml
trainingTaskInputs:
fn::toJSON:
workerPoolSpecs:
- machineSpec:
machineType: n1-standard-4
replicaCount: '1'
diskSpec:
bootDiskType: pd-ssd
bootDiskSizeGb: 100
containerSpec:
imageUri: gcr.io/cloud-marketplace/google/debian11
inputDataConfig:
datasetId: ${dataset.name}
annotationSchemaUri: gs://google-cloud-aiplatform/schema/dataset/annotation/image_classification_1.0.0.yaml
annotationsFilter: annotation_filter_key:annotation_filter_value
persistMlUseAssignment: false
filterSplit:
trainingFilter: labels.test:training
validationFilter: labels.test:validation
testFilter: labels.test:test
gcsDestination:
outputUriPrefix: gs://${bucket.name}/dataset-export
labels:
env: test
dataset:
type: gcp:vertex:AiDataset
properties:
displayName: test-dataset
metadataSchemaUri: gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml
region: us-central1
bucket:
type: gcp:storage:Bucket
properties:
name: test-bucket
location: us-central1
forceDestroy: true
uniformBucketLevelAccess: true
terraform {
required_providers {
gcp = {
source = "pulumi/gcp"
}
}
}
resource "gcp_vertex_ai_training_pipeline" "dataset_pipeline" {
lifecycle {
create_before_destroy = true
}
display_name = "test-training-pipeline-dataset"
model_id = "test-model-id"
region = "us-central1"
training_task_definition = "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml"
training_task_inputs = jsonencode({
"workerPoolSpecs" = [{
"machineSpec" = {
"machineType" = "n1-standard-4"
}
"replicaCount" = "1"
"diskSpec" = {
"bootDiskType" = "pd-ssd"
"bootDiskSizeGb" = 100
}
"containerSpec" = {
"imageUri" = "gcr.io/cloud-marketplace/google/debian11"
}
}]
})
input_data_config = {
dataset_id = gcp_vertex_ai_dataset.dataset.name
annotation_schema_uri = "gs://google-cloud-aiplatform/schema/dataset/annotation/image_classification_1.0.0.yaml"
annotations_filter = "annotation_filter_key:annotation_filter_value"
persist_ml_use_assignment = false
filter_split = {
training_filter = "labels.test:training"
validation_filter = "labels.test:validation"
test_filter = "labels.test:test"
}
gcs_destination = {
output_uri_prefix ="gs://${gcp_storage_bucket.bucket.name}/dataset-export"
}
}
labels = {
"env" = "test"
}
}
resource "gcp_vertex_ai_dataset" "dataset" {
lifecycle {
create_before_destroy = true
}
display_name = "test-dataset"
metadata_schema_uri = "gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml"
region = "us-central1"
}
resource "gcp_storage_bucket" "bucket" {
lifecycle {
create_before_destroy = true
}
name = "test-bucket"
location = "us-central1"
force_destroy = true
uniform_bucket_level_access = true
}
Vertex Ai Training Pipeline Encryption
import * as pulumi from "@pulumi/pulumi";
import * as gcp from "@pulumi/gcp";
const encryptionPipeline = new gcp.vertex.AiTrainingPipeline("encryption_pipeline", {
displayName: "test-training-pipeline-enc",
modelId: "test-model-id",
region: "us-central1",
trainingTaskDefinition: "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
trainingTaskInputs: JSON.stringify({
workerPoolSpecs: [{
machineSpec: {
machineType: "n1-standard-4",
},
replicaCount: "1",
diskSpec: {
bootDiskType: "pd-ssd",
bootDiskSizeGb: 100,
},
containerSpec: {
imageUri: "gcr.io/cloud-marketplace/google/debian11",
},
}],
}),
modelToUpload: {
displayName: "test-model-enc",
artifactUri: "gs://cloud-samples-data/ai-platform/flower_photos",
containerSpec: {
imageUri: "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest",
},
encryptionSpec: {
kmsKeyName: "kms-key",
},
},
encryptionSpec: {
kmsKeyName: "kms-key",
},
labels: {
env: "test",
},
});
import pulumi
import json
import pulumi_gcp as gcp
encryption_pipeline = gcp.vertex.AiTrainingPipeline("encryption_pipeline",
display_name="test-training-pipeline-enc",
model_id="test-model-id",
region="us-central1",
training_task_definition="gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
training_task_inputs=json.dumps({
"workerPoolSpecs": [{
"machineSpec": {
"machineType": "n1-standard-4",
},
"replicaCount": "1",
"diskSpec": {
"bootDiskType": "pd-ssd",
"bootDiskSizeGb": 100,
},
"containerSpec": {
"imageUri": "gcr.io/cloud-marketplace/google/debian11",
},
}],
}),
model_to_upload={
"display_name": "test-model-enc",
"artifact_uri": "gs://cloud-samples-data/ai-platform/flower_photos",
"container_spec": {
"image_uri": "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest",
},
"encryption_spec": {
"kms_key_name": "kms-key",
},
},
encryption_spec={
"kms_key_name": "kms-key",
},
labels={
"env": "test",
})
package main
import (
"encoding/json"
"github.com/pulumi/pulumi-gcp/sdk/v10/go/gcp/vertex"
"github.com/pulumi/pulumi/sdk/v3/go/pulumi"
)
func main() {
pulumi.Run(func(ctx *pulumi.Context) error {
tmpJSON0, err := json.Marshal(map[string][]map[string]interface{}{
"workerPoolSpecs": []map[string]interface{}{
map[string]interface{}{
"machineSpec": map[string]string{
"machineType": "n1-standard-4",
},
"replicaCount": "1",
"diskSpec": map[string]interface{}{
"bootDiskType": "pd-ssd",
"bootDiskSizeGb": 100,
},
"containerSpec": map[string]string{
"imageUri": "gcr.io/cloud-marketplace/google/debian11",
},
},
},
})
if err != nil {
return err
}
json0 := string(tmpJSON0)
_, err = vertex.NewAiTrainingPipeline(ctx, "encryption_pipeline", &vertex.AiTrainingPipelineArgs{
DisplayName: pulumi.String("test-training-pipeline-enc"),
ModelId: pulumi.String("test-model-id"),
Region: pulumi.String("us-central1"),
TrainingTaskDefinition: pulumi.String("gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml"),
TrainingTaskInputs: pulumi.String(json0),
ModelToUpload: &vertex.AiTrainingPipelineModelToUploadArgs{
DisplayName: pulumi.String("test-model-enc"),
ArtifactUri: pulumi.String("gs://cloud-samples-data/ai-platform/flower_photos"),
ContainerSpec: &vertex.AiTrainingPipelineModelToUploadContainerSpecArgs{
ImageUri: pulumi.String("us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest"),
},
EncryptionSpec: &vertex.AiTrainingPipelineModelToUploadEncryptionSpecArgs{
KmsKeyName: pulumi.String("kms-key"),
},
},
EncryptionSpec: &vertex.AiTrainingPipelineEncryptionSpecArgs{
KmsKeyName: pulumi.String("kms-key"),
},
Labels: pulumi.StringMap{
"env": pulumi.String("test"),
},
})
if err != nil {
return err
}
return nil
})
}
using System.Collections.Generic;
using System.Linq;
using System.Text.Json;
using Pulumi;
using Gcp = Pulumi.Gcp;
return await Deployment.RunAsync(() =>
{
var encryptionPipeline = new Gcp.Vertex.AiTrainingPipeline("encryption_pipeline", new()
{
DisplayName = "test-training-pipeline-enc",
ModelId = "test-model-id",
Region = "us-central1",
TrainingTaskDefinition = "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml",
TrainingTaskInputs = JsonSerializer.Serialize(new Dictionary<string, object?>
{
["workerPoolSpecs"] = new[]
{
new Dictionary<string, object?>
{
["machineSpec"] = new Dictionary<string, object?>
{
["machineType"] = "n1-standard-4",
},
["replicaCount"] = "1",
["diskSpec"] = new Dictionary<string, object?>
{
["bootDiskType"] = "pd-ssd",
["bootDiskSizeGb"] = 100,
},
["containerSpec"] = new Dictionary<string, object?>
{
["imageUri"] = "gcr.io/cloud-marketplace/google/debian11",
},
},
},
}),
ModelToUpload = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadArgs
{
DisplayName = "test-model-enc",
ArtifactUri = "gs://cloud-samples-data/ai-platform/flower_photos",
ContainerSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecArgs
{
ImageUri = "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest",
},
EncryptionSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadEncryptionSpecArgs
{
KmsKeyName = "kms-key",
},
},
EncryptionSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineEncryptionSpecArgs
{
KmsKeyName = "kms-key",
},
Labels =
{
{ "env", "test" },
},
});
});
package generated_program;
import com.pulumi.Context;
import com.pulumi.Pulumi;
import com.pulumi.core.Output;
import com.pulumi.gcp.vertex.AiTrainingPipeline;
import com.pulumi.gcp.vertex.AiTrainingPipelineArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadContainerSpecArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineModelToUploadEncryptionSpecArgs;
import com.pulumi.gcp.vertex.inputs.AiTrainingPipelineEncryptionSpecArgs;
import static com.pulumi.codegen.internal.Serialization.*;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Map;
import java.io.File;
import java.nio.file.Files;
import java.nio.file.Paths;
public class App {
public static void main(String[] args) {
Pulumi.run(App::stack);
}
public static void stack(Context ctx) {
var encryptionPipeline = new AiTrainingPipeline("encryptionPipeline", AiTrainingPipelineArgs.builder()
.displayName("test-training-pipeline-enc")
.modelId("test-model-id")
.region("us-central1")
.trainingTaskDefinition("gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml")
.trainingTaskInputs(serializeJson(
jsonObject(
jsonProperty("workerPoolSpecs", jsonArray(jsonObject(
jsonProperty("machineSpec", jsonObject(
jsonProperty("machineType", "n1-standard-4")
)),
jsonProperty("replicaCount", "1"),
jsonProperty("diskSpec", jsonObject(
jsonProperty("bootDiskType", "pd-ssd"),
jsonProperty("bootDiskSizeGb", 100)
)),
jsonProperty("containerSpec", jsonObject(
jsonProperty("imageUri", "gcr.io/cloud-marketplace/google/debian11")
))
)))
)))
.modelToUpload(AiTrainingPipelineModelToUploadArgs.builder()
.displayName("test-model-enc")
.artifactUri("gs://cloud-samples-data/ai-platform/flower_photos")
.containerSpec(AiTrainingPipelineModelToUploadContainerSpecArgs.builder()
.imageUri("us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest")
.build())
.encryptionSpec(AiTrainingPipelineModelToUploadEncryptionSpecArgs.builder()
.kmsKeyName("kms-key")
.build())
.build())
.encryptionSpec(AiTrainingPipelineEncryptionSpecArgs.builder()
.kmsKeyName("kms-key")
.build())
.labels(Map.of("env", "test"))
.build());
}
}
resources:
encryptionPipeline:
type: gcp:vertex:AiTrainingPipeline
name: encryption_pipeline
properties:
displayName: test-training-pipeline-enc
modelId: test-model-id
region: us-central1
trainingTaskDefinition: gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml
trainingTaskInputs:
fn::toJSON:
workerPoolSpecs:
- machineSpec:
machineType: n1-standard-4
replicaCount: '1'
diskSpec:
bootDiskType: pd-ssd
bootDiskSizeGb: 100
containerSpec:
imageUri: gcr.io/cloud-marketplace/google/debian11
modelToUpload:
displayName: test-model-enc
artifactUri: gs://cloud-samples-data/ai-platform/flower_photos
containerSpec:
imageUri: us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest
encryptionSpec:
kmsKeyName: kms-key
encryptionSpec:
kmsKeyName: kms-key
labels:
env: test
terraform {
required_providers {
gcp = {
source = "pulumi/gcp"
}
}
}
resource "gcp_vertex_ai_training_pipeline" "encryption_pipeline" {
lifecycle {
create_before_destroy = true
}
display_name = "test-training-pipeline-enc"
model_id = "test-model-id"
region = "us-central1"
training_task_definition = "gs://google-cloud-aiplatform/schema/trainingjob/definition/custom_task_1.0.0.yaml"
training_task_inputs = jsonencode({
"workerPoolSpecs" = [{
"machineSpec" = {
"machineType" = "n1-standard-4"
}
"replicaCount" = "1"
"diskSpec" = {
"bootDiskType" = "pd-ssd"
"bootDiskSizeGb" = 100
}
"containerSpec" = {
"imageUri" = "gcr.io/cloud-marketplace/google/debian11"
}
}]
})
model_to_upload = {
display_name = "test-model-enc"
artifact_uri = "gs://cloud-samples-data/ai-platform/flower_photos"
container_spec = {
image_uri = "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest"
}
encryption_spec = {
kms_key_name = "kms-key"
}
}
encryption_spec = {
kms_key_name = "kms-key"
}
labels = {
"env" = "test"
}
}
Create AiTrainingPipeline Resource
Resources are created with functions called constructors. To learn more about declaring and configuring resources, see Resources.
Constructor syntax
new AiTrainingPipeline(name: string, args: AiTrainingPipelineArgs, opts?: CustomResourceOptions);@overload
def AiTrainingPipeline(resource_name: str,
args: AiTrainingPipelineArgs,
opts: Optional[ResourceOptions] = None)
@overload
def AiTrainingPipeline(resource_name: str,
opts: Optional[ResourceOptions] = None,
display_name: Optional[str] = None,
model_id: Optional[str] = None,
training_task_definition: Optional[str] = None,
training_task_inputs: Optional[str] = None,
deletion_policy: Optional[str] = None,
encryption_spec: Optional[AiTrainingPipelineEncryptionSpecArgs] = None,
input_data_config: Optional[AiTrainingPipelineInputDataConfigArgs] = None,
labels: Optional[Mapping[str, str]] = None,
model_to_upload: Optional[AiTrainingPipelineModelToUploadArgs] = None,
parent_model: Optional[str] = None,
project: Optional[str] = None,
region: Optional[str] = None)func NewAiTrainingPipeline(ctx *Context, name string, args AiTrainingPipelineArgs, opts ...ResourceOption) (*AiTrainingPipeline, error)public AiTrainingPipeline(string name, AiTrainingPipelineArgs args, CustomResourceOptions? opts = null)
public AiTrainingPipeline(String name, AiTrainingPipelineArgs args)
public AiTrainingPipeline(String name, AiTrainingPipelineArgs args, CustomResourceOptions options)
type: gcp:vertex:AiTrainingPipeline
properties: # The arguments to resource properties.
options: # Bag of options to control resource's behavior.
resource "gcp_vertex_ai_training_pipeline" "name" {
# resource properties
}Parameters
- name string
- The unique name of the resource.
- args AiTrainingPipelineArgs
- The arguments to resource properties.
- opts CustomResourceOptions
- Bag of options to control resource's behavior.
- resource_name str
- The unique name of the resource.
- args AiTrainingPipelineArgs
- The arguments to resource properties.
- opts ResourceOptions
- Bag of options to control resource's behavior.
- ctx Context
- Context object for the current deployment.
- name string
- The unique name of the resource.
- args AiTrainingPipelineArgs
- The arguments to resource properties.
- opts ResourceOption
- Bag of options to control resource's behavior.
- name string
- The unique name of the resource.
- args AiTrainingPipelineArgs
- The arguments to resource properties.
- opts CustomResourceOptions
- Bag of options to control resource's behavior.
- name String
- The unique name of the resource.
- args AiTrainingPipelineArgs
- The arguments to resource properties.
- options CustomResourceOptions
- Bag of options to control resource's behavior.
Constructor example
The following reference example uses placeholder values for all input properties.
var aiTrainingPipelineResource = new Gcp.Vertex.AiTrainingPipeline("aiTrainingPipelineResource", new()
{
DisplayName = "string",
ModelId = "string",
TrainingTaskDefinition = "string",
TrainingTaskInputs = "string",
DeletionPolicy = "string",
EncryptionSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineEncryptionSpecArgs
{
KmsKeyName = "string",
},
InputDataConfig = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigArgs
{
DatasetId = "string",
AnnotationSchemaUri = "string",
AnnotationsFilter = "string",
BigqueryDestination = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigBigqueryDestinationArgs
{
OutputUri = "string",
},
FilterSplit = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigFilterSplitArgs
{
TestFilter = "string",
TrainingFilter = "string",
ValidationFilter = "string",
},
FractionSplit = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigFractionSplitArgs
{
TestFraction = 0.0,
TrainingFraction = 0.0,
ValidationFraction = 0.0,
},
GcsDestination = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigGcsDestinationArgs
{
OutputUriPrefix = "string",
},
PersistMlUseAssignment = false,
PredefinedSplit = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigPredefinedSplitArgs
{
Key = "string",
},
SavedQueryId = "string",
StratifiedSplit = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigStratifiedSplitArgs
{
Key = "string",
TestFraction = 0.0,
TrainingFraction = 0.0,
ValidationFraction = 0.0,
},
TimestampSplit = new Gcp.Vertex.Inputs.AiTrainingPipelineInputDataConfigTimestampSplitArgs
{
Key = "string",
TestFraction = 0.0,
TrainingFraction = 0.0,
ValidationFraction = 0.0,
},
},
Labels =
{
{ "string", "string" },
},
ModelToUpload = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadArgs
{
DisplayName = "string",
ArtifactUri = "string",
ContainerSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecArgs
{
ImageUri = "string",
Args = new[]
{
"string",
},
Commands = new[]
{
"string",
},
DeploymentTimeout = "string",
Envs = new[]
{
new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecEnvArgs
{
Name = "string",
Value = "string",
},
},
GrpcPorts = new[]
{
new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecGrpcPortArgs
{
ContainerPort = 0,
},
},
HealthProbe = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecHealthProbeArgs
{
Exec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecHealthProbeExecArgs
{
Commands = new[]
{
"string",
},
},
PeriodSeconds = 0,
TimeoutSeconds = 0,
},
HealthRoute = "string",
Ports = new[]
{
new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecPortArgs
{
ContainerPort = 0,
},
},
PredictRoute = "string",
SharedMemorySizeMb = "string",
StartupProbe = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecStartupProbeArgs
{
Exec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadContainerSpecStartupProbeExecArgs
{
Commands = new[]
{
"string",
},
},
PeriodSeconds = 0,
TimeoutSeconds = 0,
},
},
Description = "string",
EncryptionSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadEncryptionSpecArgs
{
KmsKeyName = "string",
},
ExplanationSpec = new Gcp.Vertex.Inputs.AiTrainingPipelineModelToUploadExplanationSpecArgs
{
Metadata = "string",
Parameters = "string",
},
Labels =
{
{ "string", "string" },
},
Name = "string",
VersionAliases = new[]
{
"string",
},
VersionDescription = "string",
},
ParentModel = "string",
Project = "string",
Region = "string",
});
example, err := vertex.NewAiTrainingPipeline(ctx, "aiTrainingPipelineResource", &vertex.AiTrainingPipelineArgs{
DisplayName: pulumi.String("string"),
ModelId: pulumi.String("string"),
TrainingTaskDefinition: pulumi.String("string"),
TrainingTaskInputs: pulumi.String("string"),
DeletionPolicy: pulumi.String("string"),
EncryptionSpec: &vertex.AiTrainingPipelineEncryptionSpecArgs{
KmsKeyName: pulumi.String("string"),
},
InputDataConfig: &vertex.AiTrainingPipelineInputDataConfigArgs{
DatasetId: pulumi.String("string"),
AnnotationSchemaUri: pulumi.String("string"),
AnnotationsFilter: pulumi.String("string"),
BigqueryDestination: &vertex.AiTrainingPipelineInputDataConfigBigqueryDestinationArgs{
OutputUri: pulumi.String("string"),
},
FilterSplit: &vertex.AiTrainingPipelineInputDataConfigFilterSplitArgs{
TestFilter: pulumi.String("string"),
TrainingFilter: pulumi.String("string"),
ValidationFilter: pulumi.String("string"),
},
FractionSplit: &vertex.AiTrainingPipelineInputDataConfigFractionSplitArgs{
TestFraction: pulumi.Float64(0),
TrainingFraction: pulumi.Float64(0),
ValidationFraction: pulumi.Float64(0),
},
GcsDestination: &vertex.AiTrainingPipelineInputDataConfigGcsDestinationArgs{
OutputUriPrefix: pulumi.String("string"),
},
PersistMlUseAssignment: pulumi.Bool(false),
PredefinedSplit: &vertex.AiTrainingPipelineInputDataConfigPredefinedSplitArgs{
Key: pulumi.String("string"),
},
SavedQueryId: pulumi.String("string"),
StratifiedSplit: &vertex.AiTrainingPipelineInputDataConfigStratifiedSplitArgs{
Key: pulumi.String("string"),
TestFraction: pulumi.Float64(0),
TrainingFraction: pulumi.Float64(0),
ValidationFraction: pulumi.Float64(0),
},
TimestampSplit: &vertex.AiTrainingPipelineInputDataConfigTimestampSplitArgs{
Key: pulumi.String("string"),
TestFraction: pulumi.Float64(0),
TrainingFraction: pulumi.Float64(0),
ValidationFraction: pulumi.Float64(0),
},
},
Labels: pulumi.StringMap{
"string": pulumi.String("string"),
},
ModelToUpload: &vertex.AiTrainingPipelineModelToUploadArgs{
DisplayName: pulumi.String("string"),
ArtifactUri: pulumi.String("string"),
ContainerSpec: &vertex.AiTrainingPipelineModelToUploadContainerSpecArgs{
ImageUri: pulumi.String("string"),
Args: pulumi.StringArray{
pulumi.String("string"),
},
Commands: pulumi.StringArray{
pulumi.String("string"),
},
DeploymentTimeout: pulumi.String("string"),
Envs: vertex.AiTrainingPipelineModelToUploadContainerSpecEnvArray{
&vertex.AiTrainingPipelineModelToUploadContainerSpecEnvArgs{
Name: pulumi.String("string"),
Value: pulumi.String("string"),
},
},
GrpcPorts: vertex.AiTrainingPipelineModelToUploadContainerSpecGrpcPortArray{
&vertex.AiTrainingPipelineModelToUploadContainerSpecGrpcPortArgs{
ContainerPort: pulumi.Int(0),
},
},
HealthProbe: &vertex.AiTrainingPipelineModelToUploadContainerSpecHealthProbeArgs{
Exec: &vertex.AiTrainingPipelineModelToUploadContainerSpecHealthProbeExecArgs{
Commands: pulumi.StringArray{
pulumi.String("string"),
},
},
PeriodSeconds: pulumi.Int(0),
TimeoutSeconds: pulumi.Int(0),
},
HealthRoute: pulumi.String("string"),
Ports: vertex.AiTrainingPipelineModelToUploadContainerSpecPortArray{
&vertex.AiTrainingPipelineModelToUploadContainerSpecPortArgs{
ContainerPort: pulumi.Int(0),
},
},
PredictRoute: pulumi.String("string"),
SharedMemorySizeMb: pulumi.String("string"),
StartupProbe: &vertex.AiTrainingPipelineModelToUploadContainerSpecStartupProbeArgs{
Exec: &vertex.AiTrainingPipelineModelToUploadContainerSpecStartupProbeExecArgs{
Commands: pulumi.StringArray{
pulumi.String("string"),
},
},
PeriodSeconds: pulumi.Int(0),
TimeoutSeconds: pulumi.Int(0),
},
},
Description: pulumi.String("string"),
EncryptionSpec: &vertex.AiTrainingPipelineModelToUploadEncryptionSpecArgs{
KmsKeyName: pulumi.String("string"),
},
ExplanationSpec: &vertex.AiTrainingPipelineModelToUploadExplanationSpecArgs{
Metadata: pulumi.String("string"),
Parameters: pulumi.String("string"),
},
Labels: pulumi.StringMap{
"string": pulumi.String("string"),
},
Name: pulumi.String("string"),
VersionAliases: pulumi.StringArray{
pulumi.String("string"),
},
VersionDescription: pulumi.String("string"),
},
ParentModel: pulumi.String("string"),
Project: pulumi.String("string"),
Region: pulumi.String("string"),
})
resource "gcp_vertex_ai_training_pipeline" "aiTrainingPipelineResource" {
lifecycle {
create_before_destroy = true
}
display_name = "string"
model_id = "string"
training_task_definition = "string"
training_task_inputs = "string"
deletion_policy = "string"
encryption_spec = {
kms_key_name = "string"
}
input_data_config = {
dataset_id = "string"
annotation_schema_uri = "string"
annotations_filter = "string"
bigquery_destination = {
output_uri = "string"
}
filter_split = {
test_filter = "string"
training_filter = "string"
validation_filter = "string"
}
fraction_split = {
test_fraction = 0
training_fraction = 0
validation_fraction = 0
}
gcs_destination = {
output_uri_prefix = "string"
}
persist_ml_use_assignment = false
predefined_split = {
key = "string"
}
saved_query_id = "string"
stratified_split = {
key = "string"
test_fraction = 0
training_fraction = 0
validation_fraction = 0
}
timestamp_split = {
key = "string"
test_fraction = 0
training_fraction = 0
validation_fraction = 0
}
}
labels = {
"string" = "string"
}
model_to_upload = {
display_name = "string"
artifact_uri = "string"
container_spec = {
image_uri = "string"
args = ["string"]
commands = ["string"]
deployment_timeout = "string"
envs = [{
name = "string"
value = "string"
}]
grpc_ports = [{
container_port = 0
}]
health_probe = {
exec = {
commands = ["string"]
}
period_seconds = 0
timeout_seconds = 0
}
health_route = "string"
ports = [{
container_port = 0
}]
predict_route = "string"
shared_memory_size_mb = "string"
startup_probe = {
exec = {
commands = ["string"]
}
period_seconds = 0
timeout_seconds = 0
}
}
description = "string"
encryption_spec = {
kms_key_name = "string"
}
explanation_spec = {
metadata = "string"
parameters = "string"
}
labels = {
"string" = "string"
}
name = "string"
version_aliases = ["string"]
version_description = "string"
}
parent_model = "string"
project = "string"
region = "string"
}
var aiTrainingPipelineResource = new AiTrainingPipeline("aiTrainingPipelineResource", AiTrainingPipelineArgs.builder()
.displayName("string")
.modelId("string")
.trainingTaskDefinition("string")
.trainingTaskInputs("string")
.deletionPolicy("string")
.encryptionSpec(AiTrainingPipelineEncryptionSpecArgs.builder()
.kmsKeyName("string")
.build())
.inputDataConfig(AiTrainingPipelineInputDataConfigArgs.builder()
.datasetId("string")
.annotationSchemaUri("string")
.annotationsFilter("string")
.bigqueryDestination(AiTrainingPipelineInputDataConfigBigqueryDestinationArgs.builder()
.outputUri("string")
.build())
.filterSplit(AiTrainingPipelineInputDataConfigFilterSplitArgs.builder()
.testFilter("string")
.trainingFilter("string")
.validationFilter("string")
.build())
.fractionSplit(AiTrainingPipelineInputDataConfigFractionSplitArgs.builder()
.testFraction(0.0)
.trainingFraction(0.0)
.validationFraction(0.0)
.build())
.gcsDestination(AiTrainingPipelineInputDataConfigGcsDestinationArgs.builder()
.outputUriPrefix("string")
.build())
.persistMlUseAssignment(false)
.predefinedSplit(AiTrainingPipelineInputDataConfigPredefinedSplitArgs.builder()
.key("string")
.build())
.savedQueryId("string")
.stratifiedSplit(AiTrainingPipelineInputDataConfigStratifiedSplitArgs.builder()
.key("string")
.testFraction(0.0)
.trainingFraction(0.0)
.validationFraction(0.0)
.build())
.timestampSplit(AiTrainingPipelineInputDataConfigTimestampSplitArgs.builder()
.key("string")
.testFraction(0.0)
.trainingFraction(0.0)
.validationFraction(0.0)
.build())
.build())
.labels(Map.of("string", "string"))
.modelToUpload(AiTrainingPipelineModelToUploadArgs.builder()
.displayName("string")
.artifactUri("string")
.containerSpec(AiTrainingPipelineModelToUploadContainerSpecArgs.builder()
.imageUri("string")
.args("string")
.commands("string")
.deploymentTimeout("string")
.envs(AiTrainingPipelineModelToUploadContainerSpecEnvArgs.builder()
.name("string")
.value("string")
.build())
.grpcPorts(AiTrainingPipelineModelToUploadContainerSpecGrpcPortArgs.builder()
.containerPort(0)
.build())
.healthProbe(AiTrainingPipelineModelToUploadContainerSpecHealthProbeArgs.builder()
.exec(AiTrainingPipelineModelToUploadContainerSpecHealthProbeExecArgs.builder()
.commands("string")
.build())
.periodSeconds(0)
.timeoutSeconds(0)
.build())
.healthRoute("string")
.ports(AiTrainingPipelineModelToUploadContainerSpecPortArgs.builder()
.containerPort(0)
.build())
.predictRoute("string")
.sharedMemorySizeMb("string")
.startupProbe(AiTrainingPipelineModelToUploadContainerSpecStartupProbeArgs.builder()
.exec(AiTrainingPipelineModelToUploadContainerSpecStartupProbeExecArgs.builder()
.commands("string")
.build())
.periodSeconds(0)
.timeoutSeconds(0)
.build())
.build())
.description("string")
.encryptionSpec(AiTrainingPipelineModelToUploadEncryptionSpecArgs.builder()
.kmsKeyName("string")
.build())
.explanationSpec(AiTrainingPipelineModelToUploadExplanationSpecArgs.builder()
.metadata("string")
.parameters("string")
.build())
.labels(Map.of("string", "string"))
.name("string")
.versionAliases("string")
.versionDescription("string")
.build())
.parentModel("string")
.project("string")
.region("string")
.build());
ai_training_pipeline_resource = gcp.vertex.AiTrainingPipeline("aiTrainingPipelineResource",
display_name="string",
model_id="string",
training_task_definition="string",
training_task_inputs="string",
deletion_policy="string",
encryption_spec={
"kms_key_name": "string",
},
input_data_config={
"dataset_id": "string",
"annotation_schema_uri": "string",
"annotations_filter": "string",
"bigquery_destination": {
"output_uri": "string",
},
"filter_split": {
"test_filter": "string",
"training_filter": "string",
"validation_filter": "string",
},
"fraction_split": {
"test_fraction": float(0),
"training_fraction": float(0),
"validation_fraction": float(0),
},
"gcs_destination": {
"output_uri_prefix": "string",
},
"persist_ml_use_assignment": False,
"predefined_split": {
"key": "string",
},
"saved_query_id": "string",
"stratified_split": {
"key": "string",
"test_fraction": float(0),
"training_fraction": float(0),
"validation_fraction": float(0),
},
"timestamp_split": {
"key": "string",
"test_fraction": float(0),
"training_fraction": float(0),
"validation_fraction": float(0),
},
},
labels={
"string": "string",
},
model_to_upload={
"display_name": "string",
"artifact_uri": "string",
"container_spec": {
"image_uri": "string",
"args": ["string"],
"commands": ["string"],
"deployment_timeout": "string",
"envs": [{
"name": "string",
"value": "string",
}],
"grpc_ports": [{
"container_port": 0,
}],
"health_probe": {
"exec_": {
"commands": ["string"],
},
"period_seconds": 0,
"timeout_seconds": 0,
},
"health_route": "string",
"ports": [{
"container_port": 0,
}],
"predict_route": "string",
"shared_memory_size_mb": "string",
"startup_probe": {
"exec_": {
"commands": ["string"],
},
"period_seconds": 0,
"timeout_seconds": 0,
},
},
"description": "string",
"encryption_spec": {
"kms_key_name": "string",
},
"explanation_spec": {
"metadata": "string",
"parameters": "string",
},
"labels": {
"string": "string",
},
"name": "string",
"version_aliases": ["string"],
"version_description": "string",
},
parent_model="string",
project="string",
region="string")
const aiTrainingPipelineResource = new gcp.vertex.AiTrainingPipeline("aiTrainingPipelineResource", {
displayName: "string",
modelId: "string",
trainingTaskDefinition: "string",
trainingTaskInputs: "string",
deletionPolicy: "string",
encryptionSpec: {
kmsKeyName: "string",
},
inputDataConfig: {
datasetId: "string",
annotationSchemaUri: "string",
annotationsFilter: "string",
bigqueryDestination: {
outputUri: "string",
},
filterSplit: {
testFilter: "string",
trainingFilter: "string",
validationFilter: "string",
},
fractionSplit: {
testFraction: 0,
trainingFraction: 0,
validationFraction: 0,
},
gcsDestination: {
outputUriPrefix: "string",
},
persistMlUseAssignment: false,
predefinedSplit: {
key: "string",
},
savedQueryId: "string",
stratifiedSplit: {
key: "string",
testFraction: 0,
trainingFraction: 0,
validationFraction: 0,
},
timestampSplit: {
key: "string",
testFraction: 0,
trainingFraction: 0,
validationFraction: 0,
},
},
labels: {
string: "string",
},
modelToUpload: {
displayName: "string",
artifactUri: "string",
containerSpec: {
imageUri: "string",
args: ["string"],
commands: ["string"],
deploymentTimeout: "string",
envs: [{
name: "string",
value: "string",
}],
grpcPorts: [{
containerPort: 0,
}],
healthProbe: {
exec: {
commands: ["string"],
},
periodSeconds: 0,
timeoutSeconds: 0,
},
healthRoute: "string",
ports: [{
containerPort: 0,
}],
predictRoute: "string",
sharedMemorySizeMb: "string",
startupProbe: {
exec: {
commands: ["string"],
},
periodSeconds: 0,
timeoutSeconds: 0,
},
},
description: "string",
encryptionSpec: {
kmsKeyName: "string",
},
explanationSpec: {
metadata: "string",
parameters: "string",
},
labels: {
string: "string",
},
name: "string",
versionAliases: ["string"],
versionDescription: "string",
},
parentModel: "string",
project: "string",
region: "string",
});
type: gcp:vertex:AiTrainingPipeline
properties:
deletionPolicy: string
displayName: string
encryptionSpec:
kmsKeyName: string
inputDataConfig:
annotationSchemaUri: string
annotationsFilter: string
bigqueryDestination:
outputUri: string
datasetId: string
filterSplit:
testFilter: string
trainingFilter: string
validationFilter: string
fractionSplit:
testFraction: 0
trainingFraction: 0
validationFraction: 0
gcsDestination:
outputUriPrefix: string
persistMlUseAssignment: false
predefinedSplit:
key: string
savedQueryId: string
stratifiedSplit:
key: string
testFraction: 0
trainingFraction: 0
validationFraction: 0
timestampSplit:
key: string
testFraction: 0
trainingFraction: 0
validationFraction: 0
labels:
string: string
modelId: string
modelToUpload:
artifactUri: string
containerSpec:
args:
- string
commands:
- string
deploymentTimeout: string
envs:
- name: string
value: string
grpcPorts:
- containerPort: 0
healthProbe:
exec:
commands:
- string
periodSeconds: 0
timeoutSeconds: 0
healthRoute: string
imageUri: string
ports:
- containerPort: 0
predictRoute: string
sharedMemorySizeMb: string
startupProbe:
exec:
commands:
- string
periodSeconds: 0
timeoutSeconds: 0
description: string
displayName: string
encryptionSpec:
kmsKeyName: string
explanationSpec:
metadata: string
parameters: string
labels:
string: string
name: string
versionAliases:
- string
versionDescription: string
parentModel: string
project: string
region: string
trainingTaskDefinition: string
trainingTaskInputs: string
AiTrainingPipeline Resource Properties
To learn more about resource properties and how to use them, see Inputs and Outputs in the Architecture and Concepts docs.
Inputs
In Python, inputs that are objects can be passed either as argument classes or as dictionary literals.
The AiTrainingPipeline resource accepts the following input properties:
- Display
Name string - The user-defined name of this TrainingPipeline.
- Model
Id string - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- Training
Task stringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- Training
Task stringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- Deletion
Policy string - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- Encryption
Spec AiTraining Pipeline Encryption Spec - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- Input
Data AiConfig Training Pipeline Input Data Config - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- Labels Dictionary<string, string>
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- Model
To AiUpload Training Pipeline Model To Upload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- Parent
Model string - When specified, the modelToUpload will become a new version of this parent_model.
- Project string
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- Region string
- The region of the training pipeline. eg us-central1
- Display
Name string - The user-defined name of this TrainingPipeline.
- Model
Id string - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- Training
Task stringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- Training
Task stringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- Deletion
Policy string - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- Encryption
Spec AiTraining Pipeline Encryption Spec Args - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- Input
Data AiConfig Training Pipeline Input Data Config Args - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- Labels map[string]string
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- Model
To AiUpload Training Pipeline Model To Upload Args - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- Parent
Model string - When specified, the modelToUpload will become a new version of this parent_model.
- Project string
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- Region string
- The region of the training pipeline. eg us-central1
- display_
name string - The user-defined name of this TrainingPipeline.
- model_
id string - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- training_
task_ stringdefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training_
task_ stringinputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- deletion_
policy string - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- encryption_
spec object - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- input_
data_ objectconfig - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels map(string)
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model_
to_ objectupload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- parent_
model string - When specified, the modelToUpload will become a new version of this parent_model.
- project string
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- region string
- The region of the training pipeline. eg us-central1
- display
Name String - The user-defined name of this TrainingPipeline.
- model
Id String - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- training
Task StringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training
Task StringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- deletion
Policy String - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- encryption
Spec AiTraining Pipeline Encryption Spec - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- input
Data AiConfig Training Pipeline Input Data Config - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels Map<String,String>
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model
To AiUpload Training Pipeline Model To Upload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- parent
Model String - When specified, the modelToUpload will become a new version of this parent_model.
- project String
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- region String
- The region of the training pipeline. eg us-central1
- display
Name string - The user-defined name of this TrainingPipeline.
- model
Id string - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- training
Task stringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training
Task stringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- deletion
Policy string - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- encryption
Spec AiTraining Pipeline Encryption Spec - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- input
Data AiConfig Training Pipeline Input Data Config - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels {[key: string]: string}
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model
To AiUpload Training Pipeline Model To Upload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- parent
Model string - When specified, the modelToUpload will become a new version of this parent_model.
- project string
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- region string
- The region of the training pipeline. eg us-central1
- display_
name str - The user-defined name of this TrainingPipeline.
- model_
id str - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- training_
task_ strdefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training_
task_ strinputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- deletion_
policy str - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- encryption_
spec AiTraining Pipeline Encryption Spec Args - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- input_
data_ Aiconfig Training Pipeline Input Data Config Args - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels Mapping[str, str]
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model_
to_ Aiupload Training Pipeline Model To Upload Args - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- parent_
model str - When specified, the modelToUpload will become a new version of this parent_model.
- project str
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- region str
- The region of the training pipeline. eg us-central1
- display
Name String - The user-defined name of this TrainingPipeline.
- model
Id String - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- training
Task StringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training
Task StringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- deletion
Policy String - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- encryption
Spec Property Map - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- input
Data Property MapConfig - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels Map<String>
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model
To Property MapUpload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- parent
Model String - When specified, the modelToUpload will become a new version of this parent_model.
- project String
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- region String
- The region of the training pipeline. eg us-central1
Outputs
All input properties are implicitly available as output properties. Additionally, the AiTrainingPipeline resource produces the following output properties:
- Create
Time string - Time when the TrainingPipeline was created.
- Effective
Labels Dictionary<string, string> - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- End
Time string - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- Errors
List<Ai
Training Pipeline Error> - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- Id string
- The provider-assigned unique ID for this managed resource.
- Name string
- The resource name of the TrainingPipeline.
- Pulumi
Labels Dictionary<string, string> - The combination of labels configured directly on the resource and default labels configured on the provider.
- Start
Time string - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- State string
- The detailed state of the pipeline.
- Training
Task stringMetadata - The metadata information as specified in the training_task_definition's metadata.
- Update
Time string - Time when the TrainingPipeline was most recently updated.
- Create
Time string - Time when the TrainingPipeline was created.
- Effective
Labels map[string]string - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- End
Time string - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- Errors
[]Ai
Training Pipeline Error - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- Id string
- The provider-assigned unique ID for this managed resource.
- Name string
- The resource name of the TrainingPipeline.
- Pulumi
Labels map[string]string - The combination of labels configured directly on the resource and default labels configured on the provider.
- Start
Time string - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- State string
- The detailed state of the pipeline.
- Training
Task stringMetadata - The metadata information as specified in the training_task_definition's metadata.
- Update
Time string - Time when the TrainingPipeline was most recently updated.
- create_
time string - Time when the TrainingPipeline was created.
- effective_
labels map(string) - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- end_
time string - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors list(object)
- Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- id string
- The provider-assigned unique ID for this managed resource.
- name string
- The resource name of the TrainingPipeline.
- pulumi_
labels map(string) - The combination of labels configured directly on the resource and default labels configured on the provider.
- start_
time string - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state string
- The detailed state of the pipeline.
- training_
task_ stringmetadata - The metadata information as specified in the training_task_definition's metadata.
- update_
time string - Time when the TrainingPipeline was most recently updated.
- create
Time String - Time when the TrainingPipeline was created.
- effective
Labels Map<String,String> - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- end
Time String - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors
List<Ai
Training Pipeline Error> - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- id String
- The provider-assigned unique ID for this managed resource.
- name String
- The resource name of the TrainingPipeline.
- pulumi
Labels Map<String,String> - The combination of labels configured directly on the resource and default labels configured on the provider.
- start
Time String - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state String
- The detailed state of the pipeline.
- training
Task StringMetadata - The metadata information as specified in the training_task_definition's metadata.
- update
Time String - Time when the TrainingPipeline was most recently updated.
- create
Time string - Time when the TrainingPipeline was created.
- effective
Labels {[key: string]: string} - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- end
Time string - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors
Ai
Training Pipeline Error[] - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- id string
- The provider-assigned unique ID for this managed resource.
- name string
- The resource name of the TrainingPipeline.
- pulumi
Labels {[key: string]: string} - The combination of labels configured directly on the resource and default labels configured on the provider.
- start
Time string - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state string
- The detailed state of the pipeline.
- training
Task stringMetadata - The metadata information as specified in the training_task_definition's metadata.
- update
Time string - Time when the TrainingPipeline was most recently updated.
- create_
time str - Time when the TrainingPipeline was created.
- effective_
labels Mapping[str, str] - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- end_
time str - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors
Sequence[Ai
Training Pipeline Error] - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- id str
- The provider-assigned unique ID for this managed resource.
- name str
- The resource name of the TrainingPipeline.
- pulumi_
labels Mapping[str, str] - The combination of labels configured directly on the resource and default labels configured on the provider.
- start_
time str - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state str
- The detailed state of the pipeline.
- training_
task_ strmetadata - The metadata information as specified in the training_task_definition's metadata.
- update_
time str - Time when the TrainingPipeline was most recently updated.
- create
Time String - Time when the TrainingPipeline was created.
- effective
Labels Map<String> - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- end
Time String - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors List<Property Map>
- Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- id String
- The provider-assigned unique ID for this managed resource.
- name String
- The resource name of the TrainingPipeline.
- pulumi
Labels Map<String> - The combination of labels configured directly on the resource and default labels configured on the provider.
- start
Time String - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state String
- The detailed state of the pipeline.
- training
Task StringMetadata - The metadata information as specified in the training_task_definition's metadata.
- update
Time String - Time when the TrainingPipeline was most recently updated.
Look up Existing AiTrainingPipeline Resource
Get an existing AiTrainingPipeline resource’s state with the given name, ID, and optional extra properties used to qualify the lookup.
public static get(name: string, id: Input<ID>, state?: AiTrainingPipelineState, opts?: CustomResourceOptions): AiTrainingPipeline@staticmethod
def get(resource_name: str,
id: str,
opts: Optional[ResourceOptions] = None,
create_time: Optional[str] = None,
deletion_policy: Optional[str] = None,
display_name: Optional[str] = None,
effective_labels: Optional[Mapping[str, str]] = None,
encryption_spec: Optional[AiTrainingPipelineEncryptionSpecArgs] = None,
end_time: Optional[str] = None,
errors: Optional[Sequence[AiTrainingPipelineErrorArgs]] = None,
input_data_config: Optional[AiTrainingPipelineInputDataConfigArgs] = None,
labels: Optional[Mapping[str, str]] = None,
model_id: Optional[str] = None,
model_to_upload: Optional[AiTrainingPipelineModelToUploadArgs] = None,
name: Optional[str] = None,
parent_model: Optional[str] = None,
project: Optional[str] = None,
pulumi_labels: Optional[Mapping[str, str]] = None,
region: Optional[str] = None,
start_time: Optional[str] = None,
state: Optional[str] = None,
training_task_definition: Optional[str] = None,
training_task_inputs: Optional[str] = None,
training_task_metadata: Optional[str] = None,
update_time: Optional[str] = None) -> AiTrainingPipelinefunc GetAiTrainingPipeline(ctx *Context, name string, id IDInput, state *AiTrainingPipelineState, opts ...ResourceOption) (*AiTrainingPipeline, error)public static AiTrainingPipeline Get(string name, Input<string> id, AiTrainingPipelineState? state, CustomResourceOptions? opts = null)public static AiTrainingPipeline get(String name, Output<String> id, AiTrainingPipelineState state, CustomResourceOptions options)resources: existingAiTrainingPipeline: type: gcp:vertex:AiTrainingPipeline get: id: ${id}import {
to = gcp_vertex_ai_training_pipeline.example
id = "${id}"
}
- name string
- The unique name of the resulting resource.
- id Input<ID>
- The unique provider ID of the resource to look up.
- state AiTrainingPipelineState
- Known property values the provider needs, along with the ID, to find the resource. Usually not required.
- opts CustomResourceOptions
- Bag of options to control resource's behavior.
- resource_name str
- The unique name of the resulting resource.
- id str
- The unique provider ID of the resource to look up.
- opts ResourceOptions
- Bag of options to control resource's behavior.
- ctx Context
- Context object for the current deployment.
- name string
- The unique name of the resulting resource.
- id IDInput
- The unique provider ID of the resource to look up.
- state AiTrainingPipelineState
- Known property values the provider needs, along with the ID, to find the resource. Usually not required.
- opts ResourceOption
- Bag of options to control resource's behavior.
- name string
- The unique name of the resulting resource.
- id Input<string>
- The unique provider ID of the resource to look up.
- state AiTrainingPipelineState
- Known property values the provider needs, along with the ID, to find the resource. Usually not required.
- opts CustomResourceOptions
- Bag of options to control resource's behavior.
- name String
- The unique name of the resulting resource.
- id Output<String>
- The unique provider ID of the resource to look up.
- state AiTrainingPipelineState
- Known property values the provider needs, along with the ID, to find the resource. Usually not required.
- options CustomResourceOptions
- Bag of options to control resource's behavior.
- Create
Time string - Time when the TrainingPipeline was created.
- Deletion
Policy string - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- Display
Name string - The user-defined name of this TrainingPipeline.
- Effective
Labels Dictionary<string, string> - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- Encryption
Spec AiTraining Pipeline Encryption Spec - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- End
Time string - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- Errors
List<Ai
Training Pipeline Error> - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- Input
Data AiConfig Training Pipeline Input Data Config - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- Labels Dictionary<string, string>
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- Model
Id string - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- Model
To AiUpload Training Pipeline Model To Upload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- Name string
- The resource name of the TrainingPipeline.
- Parent
Model string - When specified, the modelToUpload will become a new version of this parent_model.
- Project string
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- Pulumi
Labels Dictionary<string, string> - The combination of labels configured directly on the resource and default labels configured on the provider.
- Region string
- The region of the training pipeline. eg us-central1
- Start
Time string - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- State string
- The detailed state of the pipeline.
- Training
Task stringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- Training
Task stringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- Training
Task stringMetadata - The metadata information as specified in the training_task_definition's metadata.
- Update
Time string - Time when the TrainingPipeline was most recently updated.
- Create
Time string - Time when the TrainingPipeline was created.
- Deletion
Policy string - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- Display
Name string - The user-defined name of this TrainingPipeline.
- Effective
Labels map[string]string - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- Encryption
Spec AiTraining Pipeline Encryption Spec Args - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- End
Time string - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- Errors
[]Ai
Training Pipeline Error Args - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- Input
Data AiConfig Training Pipeline Input Data Config Args - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- Labels map[string]string
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- Model
Id string - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- Model
To AiUpload Training Pipeline Model To Upload Args - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- Name string
- The resource name of the TrainingPipeline.
- Parent
Model string - When specified, the modelToUpload will become a new version of this parent_model.
- Project string
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- Pulumi
Labels map[string]string - The combination of labels configured directly on the resource and default labels configured on the provider.
- Region string
- The region of the training pipeline. eg us-central1
- Start
Time string - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- State string
- The detailed state of the pipeline.
- Training
Task stringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- Training
Task stringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- Training
Task stringMetadata - The metadata information as specified in the training_task_definition's metadata.
- Update
Time string - Time when the TrainingPipeline was most recently updated.
- create_
time string - Time when the TrainingPipeline was created.
- deletion_
policy string - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- display_
name string - The user-defined name of this TrainingPipeline.
- effective_
labels map(string) - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- encryption_
spec object - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- end_
time string - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors list(object)
- Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- input_
data_ objectconfig - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels map(string)
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model_
id string - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- model_
to_ objectupload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- name string
- The resource name of the TrainingPipeline.
- parent_
model string - When specified, the modelToUpload will become a new version of this parent_model.
- project string
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- pulumi_
labels map(string) - The combination of labels configured directly on the resource and default labels configured on the provider.
- region string
- The region of the training pipeline. eg us-central1
- start_
time string - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state string
- The detailed state of the pipeline.
- training_
task_ stringdefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training_
task_ stringinputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- training_
task_ stringmetadata - The metadata information as specified in the training_task_definition's metadata.
- update_
time string - Time when the TrainingPipeline was most recently updated.
- create
Time String - Time when the TrainingPipeline was created.
- deletion
Policy String - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- display
Name String - The user-defined name of this TrainingPipeline.
- effective
Labels Map<String,String> - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- encryption
Spec AiTraining Pipeline Encryption Spec - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- end
Time String - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors
List<Ai
Training Pipeline Error> - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- input
Data AiConfig Training Pipeline Input Data Config - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels Map<String,String>
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model
Id String - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- model
To AiUpload Training Pipeline Model To Upload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- name String
- The resource name of the TrainingPipeline.
- parent
Model String - When specified, the modelToUpload will become a new version of this parent_model.
- project String
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- pulumi
Labels Map<String,String> - The combination of labels configured directly on the resource and default labels configured on the provider.
- region String
- The region of the training pipeline. eg us-central1
- start
Time String - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state String
- The detailed state of the pipeline.
- training
Task StringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training
Task StringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- training
Task StringMetadata - The metadata information as specified in the training_task_definition's metadata.
- update
Time String - Time when the TrainingPipeline was most recently updated.
- create
Time string - Time when the TrainingPipeline was created.
- deletion
Policy string - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- display
Name string - The user-defined name of this TrainingPipeline.
- effective
Labels {[key: string]: string} - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- encryption
Spec AiTraining Pipeline Encryption Spec - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- end
Time string - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors
Ai
Training Pipeline Error[] - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- input
Data AiConfig Training Pipeline Input Data Config - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels {[key: string]: string}
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model
Id string - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- model
To AiUpload Training Pipeline Model To Upload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- name string
- The resource name of the TrainingPipeline.
- parent
Model string - When specified, the modelToUpload will become a new version of this parent_model.
- project string
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- pulumi
Labels {[key: string]: string} - The combination of labels configured directly on the resource and default labels configured on the provider.
- region string
- The region of the training pipeline. eg us-central1
- start
Time string - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state string
- The detailed state of the pipeline.
- training
Task stringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training
Task stringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- training
Task stringMetadata - The metadata information as specified in the training_task_definition's metadata.
- update
Time string - Time when the TrainingPipeline was most recently updated.
- create_
time str - Time when the TrainingPipeline was created.
- deletion_
policy str - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- display_
name str - The user-defined name of this TrainingPipeline.
- effective_
labels Mapping[str, str] - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- encryption_
spec AiTraining Pipeline Encryption Spec Args - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- end_
time str - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors
Sequence[Ai
Training Pipeline Error Args] - Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- input_
data_ Aiconfig Training Pipeline Input Data Config Args - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels Mapping[str, str]
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model_
id str - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- model_
to_ Aiupload Training Pipeline Model To Upload Args - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- name str
- The resource name of the TrainingPipeline.
- parent_
model str - When specified, the modelToUpload will become a new version of this parent_model.
- project str
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- pulumi_
labels Mapping[str, str] - The combination of labels configured directly on the resource and default labels configured on the provider.
- region str
- The region of the training pipeline. eg us-central1
- start_
time str - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state str
- The detailed state of the pipeline.
- training_
task_ strdefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training_
task_ strinputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- training_
task_ strmetadata - The metadata information as specified in the training_task_definition's metadata.
- update_
time str - Time when the TrainingPipeline was most recently updated.
- create
Time String - Time when the TrainingPipeline was created.
- deletion
Policy String - Whether Terraform will be prevented from destroying the resource. Defaults to DELETE. When a 'terraform destroy' or 'pulumi up' would delete the resource, the command will fail if this field is set to "PREVENT" in Terraform state. When set to "ABANDON", the command will remove the resource from Terraform management without updating or deleting the resource in the API. When set to "DELETE", deleting the resource is allowed.
- display
Name String - The user-defined name of this TrainingPipeline.
- effective
Labels Map<String> - All of labels (key/value pairs) present on the resource in GCP, including the labels configured through Pulumi, other clients and services.
- encryption
Spec Property Map - Customer-managed encryption key spec for a TrainingPipeline. Structure is documented below.
- end
Time String - Time when the TrainingPipeline entered any of the following states: PIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, PIPELINE_STATE_CANCELLED.
- errors List<Property Map>
- Only populated when the pipeline's state is PIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED. Structure is documented below.
- input
Data Property MapConfig - Specifies Vertex AI owned input data that may be used for training the Model. Structure is documented below.
- labels Map<String>
The labels with user-defined metadata to organize TrainingPipelines.
Note: This field is non-authoritative, and will only manage the labels present in your configuration. Please refer to the field
effectiveLabelsfor all of the labels present on the resource.- model
Id String - The ID to use for the uploaded Model, which will become the final component of the model resource name.
- model
To Property MapUpload - Describes the Model that may be uploaded by this TrainingPipeline. Structure is documented below.
- name String
- The resource name of the TrainingPipeline.
- parent
Model String - When specified, the modelToUpload will become a new version of this parent_model.
- project String
- The ID of the project in which the resource belongs. If it is not provided, the provider project is used.
- pulumi
Labels Map<String> - The combination of labels configured directly on the resource and default labels configured on the provider.
- region String
- The region of the training pipeline. eg us-central1
- start
Time String - Time when the TrainingPipeline for the first time entered the PIPELINE_STATE_RUNNING state.
- state String
- The detailed state of the pipeline.
- training
Task StringDefinition - A Google Cloud Storage path to the YAML file that defines the training task which is responsible for producing the model artifact, and may also include additional auxiliary work. The definition files that can be used here are found in gs://google-cloud-aiplatform/schema/trainingjob/definition/.
- training
Task StringInputs - The training task's parameter(s), as specified in the training_task_definition's inputs. This field uses JSON data as a string.
- training
Task StringMetadata - The metadata information as specified in the training_task_definition's metadata.
- update
Time String - Time when the TrainingPipeline was most recently updated.
Supporting Types
AiTrainingPipelineEncryptionSpec, AiTrainingPipelineEncryptionSpecArgs
- Kms
Key stringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- Kms
Key stringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms_
key_ stringname - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms
Key StringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms
Key stringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms_
key_ strname - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms
Key StringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
AiTrainingPipelineError, AiTrainingPipelineErrorArgs
AiTrainingPipelineInputDataConfig, AiTrainingPipelineInputDataConfigArgs
- Dataset
Id string - The ID of the Dataset in the same Project and Location which data will be used to train the Model.
- Annotation
Schema stringUri - Applicable only to custom training with Datasets that have DataItems and Annotations. Cloud Storage URI that points to a YAML file describing the annotation schema.
- Annotations
Filter string - Applicable only to Datasets that have DataItems and Annotations. A filter on Annotations of the Dataset.
- Bigquery
Destination AiTraining Pipeline Input Data Config Bigquery Destination - The BigQuery location where the training data is to be written to. Structure is documented below.
- Filter
Split AiTraining Pipeline Input Data Config Filter Split - Assigns input data to training, validation, and test sets based on the given filters. Structure is documented below.
- Fraction
Split AiTraining Pipeline Input Data Config Fraction Split - The instructions how the input data should be split between the training, validation and test sets. Structure is documented below.
- Gcs
Destination AiTraining Pipeline Input Data Config Gcs Destination - The Cloud Storage location where the training data is to be written to. Structure is documented below.
- Persist
Ml boolUse Assignment - Whether to persist the ML use assignment to data item system labels.
- Predefined
Split AiTraining Pipeline Input Data Config Predefined Split - Assigns input data to training, validation, and test sets based on the value of a provided key. Supported only for tabular Datasets. Structure is documented below.
- Saved
Query stringId - Only applicable to Datasets that have SavedQueries. The ID of a SavedQuery under the Dataset specified by datasetId used for filtering Annotations for training.
- Stratified
Split AiTraining Pipeline Input Data Config Stratified Split - Assigns input data to the training, validation, and test sets so that the distribution of values found in the categorical column is mirrored within each split. Supported only for tabular Datasets. Structure is documented below.
- Timestamp
Split AiTraining Pipeline Input Data Config Timestamp Split - Assigns input data to training, validation, and test sets based on provided timestamps. Supported only for tabular Datasets. Structure is documented below.
- Dataset
Id string - The ID of the Dataset in the same Project and Location which data will be used to train the Model.
- Annotation
Schema stringUri - Applicable only to custom training with Datasets that have DataItems and Annotations. Cloud Storage URI that points to a YAML file describing the annotation schema.
- Annotations
Filter string - Applicable only to Datasets that have DataItems and Annotations. A filter on Annotations of the Dataset.
- Bigquery
Destination AiTraining Pipeline Input Data Config Bigquery Destination - The BigQuery location where the training data is to be written to. Structure is documented below.
- Filter
Split AiTraining Pipeline Input Data Config Filter Split - Assigns input data to training, validation, and test sets based on the given filters. Structure is documented below.
- Fraction
Split AiTraining Pipeline Input Data Config Fraction Split - The instructions how the input data should be split between the training, validation and test sets. Structure is documented below.
- Gcs
Destination AiTraining Pipeline Input Data Config Gcs Destination - The Cloud Storage location where the training data is to be written to. Structure is documented below.
- Persist
Ml boolUse Assignment - Whether to persist the ML use assignment to data item system labels.
- Predefined
Split AiTraining Pipeline Input Data Config Predefined Split - Assigns input data to training, validation, and test sets based on the value of a provided key. Supported only for tabular Datasets. Structure is documented below.
- Saved
Query stringId - Only applicable to Datasets that have SavedQueries. The ID of a SavedQuery under the Dataset specified by datasetId used for filtering Annotations for training.
- Stratified
Split AiTraining Pipeline Input Data Config Stratified Split - Assigns input data to the training, validation, and test sets so that the distribution of values found in the categorical column is mirrored within each split. Supported only for tabular Datasets. Structure is documented below.
- Timestamp
Split AiTraining Pipeline Input Data Config Timestamp Split - Assigns input data to training, validation, and test sets based on provided timestamps. Supported only for tabular Datasets. Structure is documented below.
- dataset_
id string - The ID of the Dataset in the same Project and Location which data will be used to train the Model.
- annotation_
schema_ stringuri - Applicable only to custom training with Datasets that have DataItems and Annotations. Cloud Storage URI that points to a YAML file describing the annotation schema.
- annotations_
filter string - Applicable only to Datasets that have DataItems and Annotations. A filter on Annotations of the Dataset.
- bigquery_
destination object - The BigQuery location where the training data is to be written to. Structure is documented below.
- filter_
split object - Assigns input data to training, validation, and test sets based on the given filters. Structure is documented below.
- fraction_
split object - The instructions how the input data should be split between the training, validation and test sets. Structure is documented below.
- gcs_
destination object - The Cloud Storage location where the training data is to be written to. Structure is documented below.
- persist_
ml_ booluse_ assignment - Whether to persist the ML use assignment to data item system labels.
- predefined_
split object - Assigns input data to training, validation, and test sets based on the value of a provided key. Supported only for tabular Datasets. Structure is documented below.
- saved_
query_ stringid - Only applicable to Datasets that have SavedQueries. The ID of a SavedQuery under the Dataset specified by datasetId used for filtering Annotations for training.
- stratified_
split object - Assigns input data to the training, validation, and test sets so that the distribution of values found in the categorical column is mirrored within each split. Supported only for tabular Datasets. Structure is documented below.
- timestamp_
split object - Assigns input data to training, validation, and test sets based on provided timestamps. Supported only for tabular Datasets. Structure is documented below.
- dataset
Id String - The ID of the Dataset in the same Project and Location which data will be used to train the Model.
- annotation
Schema StringUri - Applicable only to custom training with Datasets that have DataItems and Annotations. Cloud Storage URI that points to a YAML file describing the annotation schema.
- annotations
Filter String - Applicable only to Datasets that have DataItems and Annotations. A filter on Annotations of the Dataset.
- bigquery
Destination AiTraining Pipeline Input Data Config Bigquery Destination - The BigQuery location where the training data is to be written to. Structure is documented below.
- filter
Split AiTraining Pipeline Input Data Config Filter Split - Assigns input data to training, validation, and test sets based on the given filters. Structure is documented below.
- fraction
Split AiTraining Pipeline Input Data Config Fraction Split - The instructions how the input data should be split between the training, validation and test sets. Structure is documented below.
- gcs
Destination AiTraining Pipeline Input Data Config Gcs Destination - The Cloud Storage location where the training data is to be written to. Structure is documented below.
- persist
Ml BooleanUse Assignment - Whether to persist the ML use assignment to data item system labels.
- predefined
Split AiTraining Pipeline Input Data Config Predefined Split - Assigns input data to training, validation, and test sets based on the value of a provided key. Supported only for tabular Datasets. Structure is documented below.
- saved
Query StringId - Only applicable to Datasets that have SavedQueries. The ID of a SavedQuery under the Dataset specified by datasetId used for filtering Annotations for training.
- stratified
Split AiTraining Pipeline Input Data Config Stratified Split - Assigns input data to the training, validation, and test sets so that the distribution of values found in the categorical column is mirrored within each split. Supported only for tabular Datasets. Structure is documented below.
- timestamp
Split AiTraining Pipeline Input Data Config Timestamp Split - Assigns input data to training, validation, and test sets based on provided timestamps. Supported only for tabular Datasets. Structure is documented below.
- dataset
Id string - The ID of the Dataset in the same Project and Location which data will be used to train the Model.
- annotation
Schema stringUri - Applicable only to custom training with Datasets that have DataItems and Annotations. Cloud Storage URI that points to a YAML file describing the annotation schema.
- annotations
Filter string - Applicable only to Datasets that have DataItems and Annotations. A filter on Annotations of the Dataset.
- bigquery
Destination AiTraining Pipeline Input Data Config Bigquery Destination - The BigQuery location where the training data is to be written to. Structure is documented below.
- filter
Split AiTraining Pipeline Input Data Config Filter Split - Assigns input data to training, validation, and test sets based on the given filters. Structure is documented below.
- fraction
Split AiTraining Pipeline Input Data Config Fraction Split - The instructions how the input data should be split between the training, validation and test sets. Structure is documented below.
- gcs
Destination AiTraining Pipeline Input Data Config Gcs Destination - The Cloud Storage location where the training data is to be written to. Structure is documented below.
- persist
Ml booleanUse Assignment - Whether to persist the ML use assignment to data item system labels.
- predefined
Split AiTraining Pipeline Input Data Config Predefined Split - Assigns input data to training, validation, and test sets based on the value of a provided key. Supported only for tabular Datasets. Structure is documented below.
- saved
Query stringId - Only applicable to Datasets that have SavedQueries. The ID of a SavedQuery under the Dataset specified by datasetId used for filtering Annotations for training.
- stratified
Split AiTraining Pipeline Input Data Config Stratified Split - Assigns input data to the training, validation, and test sets so that the distribution of values found in the categorical column is mirrored within each split. Supported only for tabular Datasets. Structure is documented below.
- timestamp
Split AiTraining Pipeline Input Data Config Timestamp Split - Assigns input data to training, validation, and test sets based on provided timestamps. Supported only for tabular Datasets. Structure is documented below.
- dataset_
id str - The ID of the Dataset in the same Project and Location which data will be used to train the Model.
- annotation_
schema_ struri - Applicable only to custom training with Datasets that have DataItems and Annotations. Cloud Storage URI that points to a YAML file describing the annotation schema.
- annotations_
filter str - Applicable only to Datasets that have DataItems and Annotations. A filter on Annotations of the Dataset.
- bigquery_
destination AiTraining Pipeline Input Data Config Bigquery Destination - The BigQuery location where the training data is to be written to. Structure is documented below.
- filter_
split AiTraining Pipeline Input Data Config Filter Split - Assigns input data to training, validation, and test sets based on the given filters. Structure is documented below.
- fraction_
split AiTraining Pipeline Input Data Config Fraction Split - The instructions how the input data should be split between the training, validation and test sets. Structure is documented below.
- gcs_
destination AiTraining Pipeline Input Data Config Gcs Destination - The Cloud Storage location where the training data is to be written to. Structure is documented below.
- persist_
ml_ booluse_ assignment - Whether to persist the ML use assignment to data item system labels.
- predefined_
split AiTraining Pipeline Input Data Config Predefined Split - Assigns input data to training, validation, and test sets based on the value of a provided key. Supported only for tabular Datasets. Structure is documented below.
- saved_
query_ strid - Only applicable to Datasets that have SavedQueries. The ID of a SavedQuery under the Dataset specified by datasetId used for filtering Annotations for training.
- stratified_
split AiTraining Pipeline Input Data Config Stratified Split - Assigns input data to the training, validation, and test sets so that the distribution of values found in the categorical column is mirrored within each split. Supported only for tabular Datasets. Structure is documented below.
- timestamp_
split AiTraining Pipeline Input Data Config Timestamp Split - Assigns input data to training, validation, and test sets based on provided timestamps. Supported only for tabular Datasets. Structure is documented below.
- dataset
Id String - The ID of the Dataset in the same Project and Location which data will be used to train the Model.
- annotation
Schema StringUri - Applicable only to custom training with Datasets that have DataItems and Annotations. Cloud Storage URI that points to a YAML file describing the annotation schema.
- annotations
Filter String - Applicable only to Datasets that have DataItems and Annotations. A filter on Annotations of the Dataset.
- bigquery
Destination Property Map - The BigQuery location where the training data is to be written to. Structure is documented below.
- filter
Split Property Map - Assigns input data to training, validation, and test sets based on the given filters. Structure is documented below.
- fraction
Split Property Map - The instructions how the input data should be split between the training, validation and test sets. Structure is documented below.
- gcs
Destination Property Map - The Cloud Storage location where the training data is to be written to. Structure is documented below.
- persist
Ml BooleanUse Assignment - Whether to persist the ML use assignment to data item system labels.
- predefined
Split Property Map - Assigns input data to training, validation, and test sets based on the value of a provided key. Supported only for tabular Datasets. Structure is documented below.
- saved
Query StringId - Only applicable to Datasets that have SavedQueries. The ID of a SavedQuery under the Dataset specified by datasetId used for filtering Annotations for training.
- stratified
Split Property Map - Assigns input data to the training, validation, and test sets so that the distribution of values found in the categorical column is mirrored within each split. Supported only for tabular Datasets. Structure is documented below.
- timestamp
Split Property Map - Assigns input data to training, validation, and test sets based on provided timestamps. Supported only for tabular Datasets. Structure is documented below.
AiTrainingPipelineInputDataConfigBigqueryDestination, AiTrainingPipelineInputDataConfigBigqueryDestinationArgs
- Output
Uri string - The BigQuery table to which the output is to be written.
- Output
Uri string - The BigQuery table to which the output is to be written.
- output_
uri string - The BigQuery table to which the output is to be written.
- output
Uri String - The BigQuery table to which the output is to be written.
- output
Uri string - The BigQuery table to which the output is to be written.
- output_
uri str - The BigQuery table to which the output is to be written.
- output
Uri String - The BigQuery table to which the output is to be written.
AiTrainingPipelineInputDataConfigFilterSplit, AiTrainingPipelineInputDataConfigFilterSplitArgs
- Test
Filter string - A filter on DataItems of the Dataset for test set.
- Training
Filter string - A filter on DataItems of the Dataset for training set.
- Validation
Filter string - A filter on DataItems of the Dataset for validation set.
- Test
Filter string - A filter on DataItems of the Dataset for test set.
- Training
Filter string - A filter on DataItems of the Dataset for training set.
- Validation
Filter string - A filter on DataItems of the Dataset for validation set.
- test_
filter string - A filter on DataItems of the Dataset for test set.
- training_
filter string - A filter on DataItems of the Dataset for training set.
- validation_
filter string - A filter on DataItems of the Dataset for validation set.
- test
Filter String - A filter on DataItems of the Dataset for test set.
- training
Filter String - A filter on DataItems of the Dataset for training set.
- validation
Filter String - A filter on DataItems of the Dataset for validation set.
- test
Filter string - A filter on DataItems of the Dataset for test set.
- training
Filter string - A filter on DataItems of the Dataset for training set.
- validation
Filter string - A filter on DataItems of the Dataset for validation set.
- test_
filter str - A filter on DataItems of the Dataset for test set.
- training_
filter str - A filter on DataItems of the Dataset for training set.
- validation_
filter str - A filter on DataItems of the Dataset for validation set.
- test
Filter String - A filter on DataItems of the Dataset for test set.
- training
Filter String - A filter on DataItems of the Dataset for training set.
- validation
Filter String - A filter on DataItems of the Dataset for validation set.
AiTrainingPipelineInputDataConfigFractionSplit, AiTrainingPipelineInputDataConfigFractionSplitArgs
- Test
Fraction double - The fraction of the input data that is to be used to evaluate the Model.
- Training
Fraction double - The fraction of the input data that is to be used to train the Model.
- Validation
Fraction double - The fraction of the input data that is to be used to validate the Model.
- Test
Fraction float64 - The fraction of the input data that is to be used to evaluate the Model.
- Training
Fraction float64 - The fraction of the input data that is to be used to train the Model.
- Validation
Fraction float64 - The fraction of the input data that is to be used to validate the Model.
- test_
fraction number - The fraction of the input data that is to be used to evaluate the Model.
- training_
fraction number - The fraction of the input data that is to be used to train the Model.
- validation_
fraction number - The fraction of the input data that is to be used to validate the Model.
- test
Fraction Double - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction Double - The fraction of the input data that is to be used to train the Model.
- validation
Fraction Double - The fraction of the input data that is to be used to validate the Model.
- test
Fraction number - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction number - The fraction of the input data that is to be used to train the Model.
- validation
Fraction number - The fraction of the input data that is to be used to validate the Model.
- test_
fraction float - The fraction of the input data that is to be used to evaluate the Model.
- training_
fraction float - The fraction of the input data that is to be used to train the Model.
- validation_
fraction float - The fraction of the input data that is to be used to validate the Model.
- test
Fraction Number - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction Number - The fraction of the input data that is to be used to train the Model.
- validation
Fraction Number - The fraction of the input data that is to be used to validate the Model.
AiTrainingPipelineInputDataConfigGcsDestination, AiTrainingPipelineInputDataConfigGcsDestinationArgs
- Output
Uri stringPrefix - The Cloud Storage URI prefix to write to.
- Output
Uri stringPrefix - The Cloud Storage URI prefix to write to.
- output_
uri_ stringprefix - The Cloud Storage URI prefix to write to.
- output
Uri StringPrefix - The Cloud Storage URI prefix to write to.
- output
Uri stringPrefix - The Cloud Storage URI prefix to write to.
- output_
uri_ strprefix - The Cloud Storage URI prefix to write to.
- output
Uri StringPrefix - The Cloud Storage URI prefix to write to.
AiTrainingPipelineInputDataConfigPredefinedSplit, AiTrainingPipelineInputDataConfigPredefinedSplitArgs
- Key string
- The key is a name of one of the Dataset's data columns.
- Key string
- The key is a name of one of the Dataset's data columns.
- key string
- The key is a name of one of the Dataset's data columns.
- key String
- The key is a name of one of the Dataset's data columns.
- key string
- The key is a name of one of the Dataset's data columns.
- key str
- The key is a name of one of the Dataset's data columns.
- key String
- The key is a name of one of the Dataset's data columns.
AiTrainingPipelineInputDataConfigStratifiedSplit, AiTrainingPipelineInputDataConfigStratifiedSplitArgs
- Key string
- The key is a name of one of the Dataset's data columns.
- Test
Fraction double - The fraction of the input data that is to be used to evaluate the Model.
- Training
Fraction double - The fraction of the input data that is to be used to train the Model.
- Validation
Fraction double - The fraction of the input data that is to be used to validate the Model.
- Key string
- The key is a name of one of the Dataset's data columns.
- Test
Fraction float64 - The fraction of the input data that is to be used to evaluate the Model.
- Training
Fraction float64 - The fraction of the input data that is to be used to train the Model.
- Validation
Fraction float64 - The fraction of the input data that is to be used to validate the Model.
- key string
- The key is a name of one of the Dataset's data columns.
- test_
fraction number - The fraction of the input data that is to be used to evaluate the Model.
- training_
fraction number - The fraction of the input data that is to be used to train the Model.
- validation_
fraction number - The fraction of the input data that is to be used to validate the Model.
- key String
- The key is a name of one of the Dataset's data columns.
- test
Fraction Double - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction Double - The fraction of the input data that is to be used to train the Model.
- validation
Fraction Double - The fraction of the input data that is to be used to validate the Model.
- key string
- The key is a name of one of the Dataset's data columns.
- test
Fraction number - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction number - The fraction of the input data that is to be used to train the Model.
- validation
Fraction number - The fraction of the input data that is to be used to validate the Model.
- key str
- The key is a name of one of the Dataset's data columns.
- test_
fraction float - The fraction of the input data that is to be used to evaluate the Model.
- training_
fraction float - The fraction of the input data that is to be used to train the Model.
- validation_
fraction float - The fraction of the input data that is to be used to validate the Model.
- key String
- The key is a name of one of the Dataset's data columns.
- test
Fraction Number - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction Number - The fraction of the input data that is to be used to train the Model.
- validation
Fraction Number - The fraction of the input data that is to be used to validate the Model.
AiTrainingPipelineInputDataConfigTimestampSplit, AiTrainingPipelineInputDataConfigTimestampSplitArgs
- Key string
- The key is a name of one of the Dataset's data columns.
- Test
Fraction double - The fraction of the input data that is to be used to evaluate the Model.
- Training
Fraction double - The fraction of the input data that is to be used to train the Model.
- Validation
Fraction double - The fraction of the input data that is to be used to validate the Model.
- Key string
- The key is a name of one of the Dataset's data columns.
- Test
Fraction float64 - The fraction of the input data that is to be used to evaluate the Model.
- Training
Fraction float64 - The fraction of the input data that is to be used to train the Model.
- Validation
Fraction float64 - The fraction of the input data that is to be used to validate the Model.
- key string
- The key is a name of one of the Dataset's data columns.
- test_
fraction number - The fraction of the input data that is to be used to evaluate the Model.
- training_
fraction number - The fraction of the input data that is to be used to train the Model.
- validation_
fraction number - The fraction of the input data that is to be used to validate the Model.
- key String
- The key is a name of one of the Dataset's data columns.
- test
Fraction Double - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction Double - The fraction of the input data that is to be used to train the Model.
- validation
Fraction Double - The fraction of the input data that is to be used to validate the Model.
- key string
- The key is a name of one of the Dataset's data columns.
- test
Fraction number - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction number - The fraction of the input data that is to be used to train the Model.
- validation
Fraction number - The fraction of the input data that is to be used to validate the Model.
- key str
- The key is a name of one of the Dataset's data columns.
- test_
fraction float - The fraction of the input data that is to be used to evaluate the Model.
- training_
fraction float - The fraction of the input data that is to be used to train the Model.
- validation_
fraction float - The fraction of the input data that is to be used to validate the Model.
- key String
- The key is a name of one of the Dataset's data columns.
- test
Fraction Number - The fraction of the input data that is to be used to evaluate the Model.
- training
Fraction Number - The fraction of the input data that is to be used to train the Model.
- validation
Fraction Number - The fraction of the input data that is to be used to validate the Model.
AiTrainingPipelineModelToUpload, AiTrainingPipelineModelToUploadArgs
- Display
Name string - The user-defined name of the Model.
- Artifact
Uri string - The path to the directory containing the Model artifact and any of its supporting files.
- Container
Spec AiTraining Pipeline Model To Upload Container Spec - The specification of the container that is to be used when deploying this Model. Structure is documented below.
- Description string
- The description of the Model.
- Encryption
Spec AiTraining Pipeline Model To Upload Encryption Spec - Customer-managed encryption key spec for a Model. Structure is documented below.
- Explanation
Spec AiTraining Pipeline Model To Upload Explanation Spec - Specification of Model explanation. Structure is documented below.
- Labels Dictionary<string, string>
- The labels with user-defined metadata to organize your Models.
- Name string
- (Output) The resource name of the Model.
- Version
Aliases List<string> - User provided version aliases so that a model version can be referenced using alias.
- Version
Description string - The description of this version.
- Display
Name string - The user-defined name of the Model.
- Artifact
Uri string - The path to the directory containing the Model artifact and any of its supporting files.
- Container
Spec AiTraining Pipeline Model To Upload Container Spec - The specification of the container that is to be used when deploying this Model. Structure is documented below.
- Description string
- The description of the Model.
- Encryption
Spec AiTraining Pipeline Model To Upload Encryption Spec - Customer-managed encryption key spec for a Model. Structure is documented below.
- Explanation
Spec AiTraining Pipeline Model To Upload Explanation Spec - Specification of Model explanation. Structure is documented below.
- Labels map[string]string
- The labels with user-defined metadata to organize your Models.
- Name string
- (Output) The resource name of the Model.
- Version
Aliases []string - User provided version aliases so that a model version can be referenced using alias.
- Version
Description string - The description of this version.
- display_
name string - The user-defined name of the Model.
- artifact_
uri string - The path to the directory containing the Model artifact and any of its supporting files.
- container_
spec object - The specification of the container that is to be used when deploying this Model. Structure is documented below.
- description string
- The description of the Model.
- encryption_
spec object - Customer-managed encryption key spec for a Model. Structure is documented below.
- explanation_
spec object - Specification of Model explanation. Structure is documented below.
- labels map(string)
- The labels with user-defined metadata to organize your Models.
- name string
- (Output) The resource name of the Model.
- version_
aliases list(string) - User provided version aliases so that a model version can be referenced using alias.
- version_
description string - The description of this version.
- display
Name String - The user-defined name of the Model.
- artifact
Uri String - The path to the directory containing the Model artifact and any of its supporting files.
- container
Spec AiTraining Pipeline Model To Upload Container Spec - The specification of the container that is to be used when deploying this Model. Structure is documented below.
- description String
- The description of the Model.
- encryption
Spec AiTraining Pipeline Model To Upload Encryption Spec - Customer-managed encryption key spec for a Model. Structure is documented below.
- explanation
Spec AiTraining Pipeline Model To Upload Explanation Spec - Specification of Model explanation. Structure is documented below.
- labels Map<String,String>
- The labels with user-defined metadata to organize your Models.
- name String
- (Output) The resource name of the Model.
- version
Aliases List<String> - User provided version aliases so that a model version can be referenced using alias.
- version
Description String - The description of this version.
- display
Name string - The user-defined name of the Model.
- artifact
Uri string - The path to the directory containing the Model artifact and any of its supporting files.
- container
Spec AiTraining Pipeline Model To Upload Container Spec - The specification of the container that is to be used when deploying this Model. Structure is documented below.
- description string
- The description of the Model.
- encryption
Spec AiTraining Pipeline Model To Upload Encryption Spec - Customer-managed encryption key spec for a Model. Structure is documented below.
- explanation
Spec AiTraining Pipeline Model To Upload Explanation Spec - Specification of Model explanation. Structure is documented below.
- labels {[key: string]: string}
- The labels with user-defined metadata to organize your Models.
- name string
- (Output) The resource name of the Model.
- version
Aliases string[] - User provided version aliases so that a model version can be referenced using alias.
- version
Description string - The description of this version.
- display_
name str - The user-defined name of the Model.
- artifact_
uri str - The path to the directory containing the Model artifact and any of its supporting files.
- container_
spec AiTraining Pipeline Model To Upload Container Spec - The specification of the container that is to be used when deploying this Model. Structure is documented below.
- description str
- The description of the Model.
- encryption_
spec AiTraining Pipeline Model To Upload Encryption Spec - Customer-managed encryption key spec for a Model. Structure is documented below.
- explanation_
spec AiTraining Pipeline Model To Upload Explanation Spec - Specification of Model explanation. Structure is documented below.
- labels Mapping[str, str]
- The labels with user-defined metadata to organize your Models.
- name str
- (Output) The resource name of the Model.
- version_
aliases Sequence[str] - User provided version aliases so that a model version can be referenced using alias.
- version_
description str - The description of this version.
- display
Name String - The user-defined name of the Model.
- artifact
Uri String - The path to the directory containing the Model artifact and any of its supporting files.
- container
Spec Property Map - The specification of the container that is to be used when deploying this Model. Structure is documented below.
- description String
- The description of the Model.
- encryption
Spec Property Map - Customer-managed encryption key spec for a Model. Structure is documented below.
- explanation
Spec Property Map - Specification of Model explanation. Structure is documented below.
- labels Map<String>
- The labels with user-defined metadata to organize your Models.
- name String
- (Output) The resource name of the Model.
- version
Aliases List<String> - User provided version aliases so that a model version can be referenced using alias.
- version
Description String - The description of this version.
AiTrainingPipelineModelToUploadContainerSpec, AiTrainingPipelineModelToUploadContainerSpecArgs
- Image
Uri string - URI of the Docker image to be used as the custom container for serving predictions.
- Args List<string>
- Specifies the arguments to the command when the container starts.
- Commands List<string>
- Specifies the command that runs when the container starts.
- Deployment
Timeout string - Deployment timeout. Limit for deployment timeout is 2 hours.
- Envs
List<Ai
Training Pipeline Model To Upload Container Spec Env> - List of environment variables to be set in the container. Structure is documented below.
- Grpc
Ports List<AiTraining Pipeline Model To Upload Container Spec Grpc Port> - List of ports to expose from the container for gRPC. Structure is documented below.
- Health
Probe AiTraining Pipeline Model To Upload Container Spec Health Probe - Specification for Kubernetes readiness probe. Structure is documented below.
- Health
Route string - HTTP path on the container to send health checks to.
- Ports
List<Ai
Training Pipeline Model To Upload Container Spec Port> - List of ports that are exposed by the container's process. Structure is documented below.
- Predict
Route string - HTTP path on the container to send prediction requests to.
- string
- The amount of the VM memory to reserve as the shared memory for the model in megabytes.
- Startup
Probe AiTraining Pipeline Model To Upload Container Spec Startup Probe - Specification for Kubernetes startup probe. Structure is documented below.
- Image
Uri string - URI of the Docker image to be used as the custom container for serving predictions.
- Args []string
- Specifies the arguments to the command when the container starts.
- Commands []string
- Specifies the command that runs when the container starts.
- Deployment
Timeout string - Deployment timeout. Limit for deployment timeout is 2 hours.
- Envs
[]Ai
Training Pipeline Model To Upload Container Spec Env - List of environment variables to be set in the container. Structure is documented below.
- Grpc
Ports []AiTraining Pipeline Model To Upload Container Spec Grpc Port - List of ports to expose from the container for gRPC. Structure is documented below.
- Health
Probe AiTraining Pipeline Model To Upload Container Spec Health Probe - Specification for Kubernetes readiness probe. Structure is documented below.
- Health
Route string - HTTP path on the container to send health checks to.
- Ports
[]Ai
Training Pipeline Model To Upload Container Spec Port - List of ports that are exposed by the container's process. Structure is documented below.
- Predict
Route string - HTTP path on the container to send prediction requests to.
- string
- The amount of the VM memory to reserve as the shared memory for the model in megabytes.
- Startup
Probe AiTraining Pipeline Model To Upload Container Spec Startup Probe - Specification for Kubernetes startup probe. Structure is documented below.
- image_
uri string - URI of the Docker image to be used as the custom container for serving predictions.
- args list(string)
- Specifies the arguments to the command when the container starts.
- commands list(string)
- Specifies the command that runs when the container starts.
- deployment_
timeout string - Deployment timeout. Limit for deployment timeout is 2 hours.
- envs list(object)
- List of environment variables to be set in the container. Structure is documented below.
- grpc_
ports list(object) - List of ports to expose from the container for gRPC. Structure is documented below.
- health_
probe object - Specification for Kubernetes readiness probe. Structure is documented below.
- health_
route string - HTTP path on the container to send health checks to.
- ports list(object)
- List of ports that are exposed by the container's process. Structure is documented below.
- predict_
route string - HTTP path on the container to send prediction requests to.
- string
- The amount of the VM memory to reserve as the shared memory for the model in megabytes.
- startup_
probe object - Specification for Kubernetes startup probe. Structure is documented below.
- image
Uri String - URI of the Docker image to be used as the custom container for serving predictions.
- args List<String>
- Specifies the arguments to the command when the container starts.
- commands List<String>
- Specifies the command that runs when the container starts.
- deployment
Timeout String - Deployment timeout. Limit for deployment timeout is 2 hours.
- envs
List<Ai
Training Pipeline Model To Upload Container Spec Env> - List of environment variables to be set in the container. Structure is documented below.
- grpc
Ports List<AiTraining Pipeline Model To Upload Container Spec Grpc Port> - List of ports to expose from the container for gRPC. Structure is documented below.
- health
Probe AiTraining Pipeline Model To Upload Container Spec Health Probe - Specification for Kubernetes readiness probe. Structure is documented below.
- health
Route String - HTTP path on the container to send health checks to.
- ports
List<Ai
Training Pipeline Model To Upload Container Spec Port> - List of ports that are exposed by the container's process. Structure is documented below.
- predict
Route String - HTTP path on the container to send prediction requests to.
- String
- The amount of the VM memory to reserve as the shared memory for the model in megabytes.
- startup
Probe AiTraining Pipeline Model To Upload Container Spec Startup Probe - Specification for Kubernetes startup probe. Structure is documented below.
- image
Uri string - URI of the Docker image to be used as the custom container for serving predictions.
- args string[]
- Specifies the arguments to the command when the container starts.
- commands string[]
- Specifies the command that runs when the container starts.
- deployment
Timeout string - Deployment timeout. Limit for deployment timeout is 2 hours.
- envs
Ai
Training Pipeline Model To Upload Container Spec Env[] - List of environment variables to be set in the container. Structure is documented below.
- grpc
Ports AiTraining Pipeline Model To Upload Container Spec Grpc Port[] - List of ports to expose from the container for gRPC. Structure is documented below.
- health
Probe AiTraining Pipeline Model To Upload Container Spec Health Probe - Specification for Kubernetes readiness probe. Structure is documented below.
- health
Route string - HTTP path on the container to send health checks to.
- ports
Ai
Training Pipeline Model To Upload Container Spec Port[] - List of ports that are exposed by the container's process. Structure is documented below.
- predict
Route string - HTTP path on the container to send prediction requests to.
- string
- The amount of the VM memory to reserve as the shared memory for the model in megabytes.
- startup
Probe AiTraining Pipeline Model To Upload Container Spec Startup Probe - Specification for Kubernetes startup probe. Structure is documented below.
- image_
uri str - URI of the Docker image to be used as the custom container for serving predictions.
- args Sequence[str]
- Specifies the arguments to the command when the container starts.
- commands Sequence[str]
- Specifies the command that runs when the container starts.
- deployment_
timeout str - Deployment timeout. Limit for deployment timeout is 2 hours.
- envs
Sequence[Ai
Training Pipeline Model To Upload Container Spec Env] - List of environment variables to be set in the container. Structure is documented below.
- grpc_
ports Sequence[AiTraining Pipeline Model To Upload Container Spec Grpc Port] - List of ports to expose from the container for gRPC. Structure is documented below.
- health_
probe AiTraining Pipeline Model To Upload Container Spec Health Probe - Specification for Kubernetes readiness probe. Structure is documented below.
- health_
route str - HTTP path on the container to send health checks to.
- ports
Sequence[Ai
Training Pipeline Model To Upload Container Spec Port] - List of ports that are exposed by the container's process. Structure is documented below.
- predict_
route str - HTTP path on the container to send prediction requests to.
- str
- The amount of the VM memory to reserve as the shared memory for the model in megabytes.
- startup_
probe AiTraining Pipeline Model To Upload Container Spec Startup Probe - Specification for Kubernetes startup probe. Structure is documented below.
- image
Uri String - URI of the Docker image to be used as the custom container for serving predictions.
- args List<String>
- Specifies the arguments to the command when the container starts.
- commands List<String>
- Specifies the command that runs when the container starts.
- deployment
Timeout String - Deployment timeout. Limit for deployment timeout is 2 hours.
- envs List<Property Map>
- List of environment variables to be set in the container. Structure is documented below.
- grpc
Ports List<Property Map> - List of ports to expose from the container for gRPC. Structure is documented below.
- health
Probe Property Map - Specification for Kubernetes readiness probe. Structure is documented below.
- health
Route String - HTTP path on the container to send health checks to.
- ports List<Property Map>
- List of ports that are exposed by the container's process. Structure is documented below.
- predict
Route String - HTTP path on the container to send prediction requests to.
- String
- The amount of the VM memory to reserve as the shared memory for the model in megabytes.
- startup
Probe Property Map - Specification for Kubernetes startup probe. Structure is documented below.
AiTrainingPipelineModelToUploadContainerSpecEnv, AiTrainingPipelineModelToUploadContainerSpecEnvArgs
AiTrainingPipelineModelToUploadContainerSpecGrpcPort, AiTrainingPipelineModelToUploadContainerSpecGrpcPortArgs
- Container
Port int - The number of the port on the container.
- Container
Port int - The number of the port on the container.
- container_
port number - The number of the port on the container.
- container
Port Integer - The number of the port on the container.
- container
Port number - The number of the port on the container.
- container_
port int - The number of the port on the container.
- container
Port Number - The number of the port on the container.
AiTrainingPipelineModelToUploadContainerSpecHealthProbe, AiTrainingPipelineModelToUploadContainerSpecHealthProbeArgs
- Exec
Ai
Training Pipeline Model To Upload Container Spec Health Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- Period
Seconds int - How often (in seconds) to perform the probe.
- Timeout
Seconds int - Number of seconds after which the probe times out.
- Exec
Ai
Training Pipeline Model To Upload Container Spec Health Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- Period
Seconds int - How often (in seconds) to perform the probe.
- Timeout
Seconds int - Number of seconds after which the probe times out.
- exec object
- ExecAction specifies a command to execute. Structure is documented below.
- period_
seconds number - How often (in seconds) to perform the probe.
- timeout_
seconds number - Number of seconds after which the probe times out.
- exec
Ai
Training Pipeline Model To Upload Container Spec Health Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- period
Seconds Integer - How often (in seconds) to perform the probe.
- timeout
Seconds Integer - Number of seconds after which the probe times out.
- exec
Ai
Training Pipeline Model To Upload Container Spec Health Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- period
Seconds number - How often (in seconds) to perform the probe.
- timeout
Seconds number - Number of seconds after which the probe times out.
- exec_
Ai
Training Pipeline Model To Upload Container Spec Health Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- period_
seconds int - How often (in seconds) to perform the probe.
- timeout_
seconds int - Number of seconds after which the probe times out.
- exec Property Map
- ExecAction specifies a command to execute. Structure is documented below.
- period
Seconds Number - How often (in seconds) to perform the probe.
- timeout
Seconds Number - Number of seconds after which the probe times out.
AiTrainingPipelineModelToUploadContainerSpecHealthProbeExec, AiTrainingPipelineModelToUploadContainerSpecHealthProbeExecArgs
- Commands List<string>
- Command is the command line to execute inside the container.
- Commands []string
- Command is the command line to execute inside the container.
- commands list(string)
- Command is the command line to execute inside the container.
- commands List<String>
- Command is the command line to execute inside the container.
- commands string[]
- Command is the command line to execute inside the container.
- commands Sequence[str]
- Command is the command line to execute inside the container.
- commands List<String>
- Command is the command line to execute inside the container.
AiTrainingPipelineModelToUploadContainerSpecPort, AiTrainingPipelineModelToUploadContainerSpecPortArgs
- Container
Port int - The number of the port on the container.
- Container
Port int - The number of the port on the container.
- container_
port number - The number of the port on the container.
- container
Port Integer - The number of the port on the container.
- container
Port number - The number of the port on the container.
- container_
port int - The number of the port on the container.
- container
Port Number - The number of the port on the container.
AiTrainingPipelineModelToUploadContainerSpecStartupProbe, AiTrainingPipelineModelToUploadContainerSpecStartupProbeArgs
- Exec
Ai
Training Pipeline Model To Upload Container Spec Startup Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- Period
Seconds int - How often (in seconds) to perform the probe.
- Timeout
Seconds int - Number of seconds after which the probe times out.
- Exec
Ai
Training Pipeline Model To Upload Container Spec Startup Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- Period
Seconds int - How often (in seconds) to perform the probe.
- Timeout
Seconds int - Number of seconds after which the probe times out.
- exec object
- ExecAction specifies a command to execute. Structure is documented below.
- period_
seconds number - How often (in seconds) to perform the probe.
- timeout_
seconds number - Number of seconds after which the probe times out.
- exec
Ai
Training Pipeline Model To Upload Container Spec Startup Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- period
Seconds Integer - How often (in seconds) to perform the probe.
- timeout
Seconds Integer - Number of seconds after which the probe times out.
- exec
Ai
Training Pipeline Model To Upload Container Spec Startup Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- period
Seconds number - How often (in seconds) to perform the probe.
- timeout
Seconds number - Number of seconds after which the probe times out.
- exec_
Ai
Training Pipeline Model To Upload Container Spec Startup Probe Exec - ExecAction specifies a command to execute. Structure is documented below.
- period_
seconds int - How often (in seconds) to perform the probe.
- timeout_
seconds int - Number of seconds after which the probe times out.
- exec Property Map
- ExecAction specifies a command to execute. Structure is documented below.
- period
Seconds Number - How often (in seconds) to perform the probe.
- timeout
Seconds Number - Number of seconds after which the probe times out.
AiTrainingPipelineModelToUploadContainerSpecStartupProbeExec, AiTrainingPipelineModelToUploadContainerSpecStartupProbeExecArgs
- Commands List<string>
- Command is the command line to execute inside the container.
- Commands []string
- Command is the command line to execute inside the container.
- commands list(string)
- Command is the command line to execute inside the container.
- commands List<String>
- Command is the command line to execute inside the container.
- commands string[]
- Command is the command line to execute inside the container.
- commands Sequence[str]
- Command is the command line to execute inside the container.
- commands List<String>
- Command is the command line to execute inside the container.
AiTrainingPipelineModelToUploadEncryptionSpec, AiTrainingPipelineModelToUploadEncryptionSpecArgs
- Kms
Key stringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- Kms
Key stringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms_
key_ stringname - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms
Key StringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms
Key stringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms_
key_ strname - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
- kms
Key StringName - The Cloud KMS resource identifier of the customer-managed encryption key used to protect the training pipeline.
AiTrainingPipelineModelToUploadExplanationSpec, AiTrainingPipelineModelToUploadExplanationSpecArgs
- Metadata string
- Metadata describing the Model's input and output for explanation. This field uses JSON data as a string.
- Parameters string
- Parameters that configure explaining of the Model's predictions. This field uses JSON data as a string.
- Metadata string
- Metadata describing the Model's input and output for explanation. This field uses JSON data as a string.
- Parameters string
- Parameters that configure explaining of the Model's predictions. This field uses JSON data as a string.
- metadata string
- Metadata describing the Model's input and output for explanation. This field uses JSON data as a string.
- parameters string
- Parameters that configure explaining of the Model's predictions. This field uses JSON data as a string.
- metadata String
- Metadata describing the Model's input and output for explanation. This field uses JSON data as a string.
- parameters String
- Parameters that configure explaining of the Model's predictions. This field uses JSON data as a string.
- metadata string
- Metadata describing the Model's input and output for explanation. This field uses JSON data as a string.
- parameters string
- Parameters that configure explaining of the Model's predictions. This field uses JSON data as a string.
- metadata str
- Metadata describing the Model's input and output for explanation. This field uses JSON data as a string.
- parameters str
- Parameters that configure explaining of the Model's predictions. This field uses JSON data as a string.
- metadata String
- Metadata describing the Model's input and output for explanation. This field uses JSON data as a string.
- parameters String
- Parameters that configure explaining of the Model's predictions. This field uses JSON data as a string.
Import
TrainingPipeline can be imported using any of these accepted formats:
projects/{{project}}/locations/{{region}}/trainingPipelines/{{name}}{{project}}/{{region}}/{{name}}{{region}}/{{name}}{{name}}
When using the pulumi import command, TrainingPipeline can be imported using one of the formats above. For example:
$ pulumi import gcp:vertex/aiTrainingPipeline:AiTrainingPipeline default projects/{{project}}/locations/{{region}}/trainingPipelines/{{name}}
$ pulumi import gcp:vertex/aiTrainingPipeline:AiTrainingPipeline default {{project}}/{{region}}/{{name}}
$ pulumi import gcp:vertex/aiTrainingPipeline:AiTrainingPipeline default {{region}}/{{name}}
$ pulumi import gcp:vertex/aiTrainingPipeline:AiTrainingPipeline default {{name}}
To learn more about importing existing cloud resources, see Importing resources.
Package Details
- Repository
- Google Cloud (GCP) pulumi/pulumi-gcp
- License
- Apache-2.0
- Notes
- This Pulumi package is based on the
google-betaTerraform Provider.
published on Wednesday, Nov 18, 2026 by Pulumi