Receber previsões em lote para o Gemini

Com as previsões em lote, é possível enviar um grande número de solicitações multimodais em uma única solicitação em lote.

Saiba mais sobre o fluxo de trabalho em lote e como formatar a entrada dados, consulte Receba previsões em lote para o Gemini.

Modelos com suporte:

Modelo Versão
Gemini 1.5 Flash gemini-1.5-flash-002
gemini-1.5-flash-001
Gemini 1.5 Pro gemini-1.5-pro-002
gemini-1.5-pro-001
Gemini 1.0 Pro gemini-1.0-pro-001
gemini-1.0-pro-002

Exemplo de sintaxe

A sintaxe a seguir mostra como enviar uma solicitação de API de previsão em lote usando o comando curl. Este exemplo é específico para o armazenamento do BigQuery.

curl -X POST \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -H "Content-Type: application/json" \
  https://${LOCATION}-aiplatform.googleapis.com/v1/projects/${PROJECT_ID}/locations/${LOCATION}/batchPredictionJobs \
  -d '{
      "displayName": "...",
      "model": "publishers/google/models/${MODEL_ID}",
      "inputConfig": {
        "instancesFormat": "bigquery",
        "bigquerySource": {
          "inputUri" : "..."
        }
      },
      "outputConfig": {
        "predictionsFormat": "bigquery",
        "bigqueryDestination": {
          "outputUri": "..."
        }
      }
  }'

Parâmetros

Confira exemplos para detalhes de implementação.

Solicitação de corpo

Parâmetros

displayName

Um nome que você escolhe para o job.

model

O modelo a ser usado na previsão em lote.

inputConfig

O formato de dados. Para a previsão em lote do Gemini, as origens de entrada do Cloud Storage e do BigQuery são aceitas.

outputConfig

A configuração de saída que determina o local de saída do modelo. Os locais de saída do Cloud Storage e do BigQuery são compatíveis.

inputConfig

Parâmetros

instancesFormat

O formato de entrada do comando. Use jsonl para o Cloud Storage ou bigquery para o BigQuery.

gcsSource.uris

O URI da origem da entrada. É um local do Cloud Storage do arquivo JSONL no formato gs://bucketname/path/to/file.jsonl.

bigquerySource.inputUri

O URI da origem da entrada. Este é um URI de tabela do BigQuery no no formato bq://project_id.dataset.table. A região do conjunto de dados de entrada do BigQuery precisa ser a mesma do job de previsão em lote da Vertex AI.

outputConfig

Parâmetros

predictionsFormat

O formato de saída da previsão. Use bigquery.

gcsDestination.outputUriPrefix

O bucket e o local do diretório do Cloud Storage, no formato gs://mybucket/path/to/output.

bigqueryDestination.outputUri

O URI do BigQuery da tabela de saída de destino, na formulário bq://project_id.dataset.table. Se a tabela já existir, ele será criado para você. A região do conjunto de dados de saída do BigQuery precisa ser a mesma do job de previsão em lote da Vertex AI.

Exemplos

Solicitar uma resposta em lote

As solicitações em lote para modelos multimodais aceitam origens de armazenamento do Cloud Storage e do BigQuery. Para saber mais, consulte:

Dependendo do número de itens de entrada enviados, uma tarefa de previsão em lote pode levar algum tempo para ser concluída.

REST

Para criar um job de previsão em lote, use o método projects.locations.batchPredictionJobs.create.

Entrada do Cloud Storage

Antes de usar os dados da solicitação abaixo, faça as substituições a seguir:

  • LOCATION: uma região compatível com modelos Gemini.
  • PROJECT_ID: o ID do projeto.
  • INPUT_URI: o local do Cloud Storage da entrada de previsão em lote JSONL, como gs://bucketname/path/to/file.jsonl.
  • OUTPUT_FORMAT: para gerar saída em uma tabela do BigQuery, especifique bigquery. Para gerar saída em um bucket do Cloud Storage, especifique jsonl.
  • DESTINATION: para o BigQuery, especifique bigqueryDestination. Para o Cloud Storage, especifique gcsDestination.
  • OUTPUT_URI_FIELD_NAME: para o BigQuery, especifique outputUri. Para o Cloud Storage, especifique outputUriPrefix.
  • OUTPUT_URI: para o BigQuery, especifique o local da tabela, como bq://myproject.mydataset.output_result. A região do conjunto de dados de saída do BigQuery precisa ser a mesma do job de previsão em lote da Vertex AI. Para o Cloud Storage, especifique o bucket e o local do diretório, como gs://mybucket/path/to/output.

Corpo JSON da solicitação:

{
  "displayName": "my-cloud-storage-batch-prediction-job",
  "model": "publishers/google/models/gemini-1.5-flash-002",
  "inputConfig": {
    "instancesFormat": "jsonl",
    "gcsSource": {
      "uris" : "INPUT_URI"
    }
  },
  "outputConfig": {
    "predictionsFormat": "OUTPUT_FORMAT",
    "DESTINATION": {
      "OUTPUT_URI_FIELD_NAME": "OUTPUT_URI"
    }
  }
}

Para enviar a solicitação, escolha uma destas opções:

curl

Salve o corpo da solicitação em um arquivo com o nome request.json e execute o comando a seguir:

curl -X POST \
-H "Authorization: Bearer $(gcloud auth print-access-token)" \
-H "Content-Type: application/json; charset=utf-8" \
-d @request.json \
"https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/batchPredictionJobs"

PowerShell

Salve o corpo da solicitação em um arquivo com o nome request.json e execute o comando a seguir:

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
-Method POST `
-Headers $headers `
-ContentType: "application/json; charset=utf-8" `
-InFile request.json `
-Uri "https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/batchPredictionJobs" | Select-Object -Expand Content

Você receberá uma resposta JSON semelhante a seguinte.

Entrada do BigQuery

Antes de usar os dados da solicitação abaixo, faça as substituições a seguir:

  • LOCATION: uma região compatível com modelos Gemini.
  • PROJECT_ID: o ID do projeto.
  • INPUT_URI: a tabela do BigQuery em que a entrada de previsão em lote está localizada, como bq://myproject.mydataset.input_table. Não há suporte para conjuntos de dados multirregionais.
  • OUTPUT_FORMAT: para gerar saída em uma tabela do BigQuery, especifique bigquery. Para gerar saída em um bucket do Cloud Storage, especifique jsonl.
  • DESTINATION: para o BigQuery, especifique bigqueryDestination. Para o Cloud Storage, especifique gcsDestination.
  • OUTPUT_URI_FIELD_NAME: para o BigQuery, especifique outputUri. Para o Cloud Storage, especifique outputUriPrefix.
  • OUTPUT_URI: para o BigQuery, especifique o local da tabela, como bq://myproject.mydataset.output_result. A região do conjunto de dados de saída do BigQuery precisa ser a mesma do job de previsão em lote da Vertex AI. Para o Cloud Storage, especifique o bucket e o local do diretório, como gs://mybucket/path/to/output.

Corpo JSON da solicitação:

{
  "displayName": "my-bigquery-batch-prediction-job",
  "model": "publishers/google/models/gemini-1.5-flash-002",
  "inputConfig": {
    "instancesFormat": "bigquery",
    "bigquerySource":{
      "inputUri" : "INPUT_URI"
    }
  },
  "outputConfig": {
    "predictionsFormat": "OUTPUT_FORMAT",
    "DESTINATION": {
      "OUTPUT_URI_FIELD_NAME": "OUTPUT_URI"
    }
  }
}

Para enviar a solicitação, escolha uma destas opções:

curl

Salve o corpo da solicitação em um arquivo com o nome request.json e execute o comando a seguir:

curl -X POST \
-H "Authorization: Bearer $(gcloud auth print-access-token)" \
-H "Content-Type: application/json; charset=utf-8" \
-d @request.json \
"https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/batchPredictionJobs"

PowerShell

Salve o corpo da solicitação em um arquivo com o nome request.json e execute o comando a seguir:

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
-Method POST `
-Headers $headers `
-ContentType: "application/json; charset=utf-8" `
-InFile request.json `
-Uri "https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/batchPredictionJobs" | Select-Object -Expand Content

Você receberá uma resposta JSON semelhante a seguinte.

A resposta inclui um identificador exclusivo para a tarefa em lote. É possível pesquisar o status da tarefa em lote usando BATCH_JOB_ID até que o job state seja JOB_STATE_SUCCEEDED. Exemplo:

curl -X GET \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -H "Content-Type: application/json" \
https://us-central1-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/us-central1/batchPredictionJobs/BATCH_JOB_ID

Python

Para saber como instalar o SDK da Vertex AI para Python, consulte Instalar o SDK da Vertex AI para Python. Para mais informações, consulte a documentação de referência da API Python.

Entrada do Cloud Storage

import time
import vertexai

from vertexai.batch_prediction import BatchPredictionJob

# TODO(developer): Update and un-comment below line
# PROJECT_ID = "your-project-id"

# Initialize vertexai
vertexai.init(project=PROJECT_ID, location="us-central1")

input_uri = "gs://cloud-samples-data/batch/prompt_for_batch_gemini_predict.jsonl"

# Submit a batch prediction job with Gemini model
batch_prediction_job = BatchPredictionJob.submit(
    source_model="gemini-1.5-flash-002",
    input_dataset=input_uri,
    output_uri_prefix=output_uri,
)

# Check job status
print(f"Job resource name: {batch_prediction_job.resource_name}")
print(f"Model resource name with the job: {batch_prediction_job.model_name}")
print(f"Job state: {batch_prediction_job.state.name}")

# Refresh the job until complete
while not batch_prediction_job.has_ended:
    time.sleep(5)
    batch_prediction_job.refresh()

# Check if the job succeeds
if batch_prediction_job.has_succeeded:
    print("Job succeeded!")
else:
    print(f"Job failed: {batch_prediction_job.error}")

# Check the location of the output
print(f"Job output location: {batch_prediction_job.output_location}")

# Example response:
#  Job output location: gs://your-bucket/gen-ai-batch-prediction/prediction-model-year-month-day-hour:minute:second.12345

Entrada do BigQuery

import time
import vertexai

from vertexai.batch_prediction import BatchPredictionJob

# TODO(developer): Update and un-comment below line
# PROJECT_ID = "your-project-id"

# Initialize vertexai
vertexai.init(project=PROJECT_ID, location="us-central1")

input_uri = "bq://storage-samples.generative_ai.batch_requests_for_multimodal_input"

# Submit a batch prediction job with Gemini model
batch_prediction_job = BatchPredictionJob.submit(
    source_model="gemini-1.5-flash-002",
    input_dataset=input_uri,
    output_uri_prefix=output_uri,
)

# Check job status
print(f"Job resource name: {batch_prediction_job.resource_name}")
print(f"Model resource name with the job: {batch_prediction_job.model_name}")
print(f"Job state: {batch_prediction_job.state.name}")

# Refresh the job until complete
while not batch_prediction_job.has_ended:
    time.sleep(5)
    batch_prediction_job.refresh()

# Check if the job succeeds
if batch_prediction_job.has_succeeded:
    print("Job succeeded!")
else:
    print(f"Job failed: {batch_prediction_job.error}")

# Check the location of the output
print(f"Job output location: {batch_prediction_job.output_location}")

# Example response:
#  Job output location: bq://Project-ID/gen-ai-batch-prediction/predictions-model-year-month-day-hour:minute:second.12345

Node.js

Antes de testar esse exemplo, siga as instruções de configuração para Node.js no Guia de início rápido da Vertex AI sobre como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Vertex AI para Node.js.

Para autenticar na Vertex AI, configure o Application Default Credentials. Para mais informações, consulte Configurar a autenticação para um ambiente de desenvolvimento local.

Entrada do Cloud Storage

// Import the aiplatform library
const aiplatformLib = require('@google-cloud/aiplatform');
const aiplatform = aiplatformLib.protos.google.cloud.aiplatform.v1;

/**
 * TODO(developer):  Uncomment/update these variables before running the sample.
 */
// projectId = 'YOUR_PROJECT_ID';
// URI of the output folder in Google Cloud Storage.
// E.g. "gs://[BUCKET]/[OUTPUT]"
// outputUri = 'gs://my-bucket';

// URI of the input file in Google Cloud Storage.
// E.g. "gs://[BUCKET]/[DATASET].jsonl"
// Or try:
// "gs://cloud-samples-data/generative-ai/batch/gemini_multimodal_batch_predict.jsonl"
// for a batch prediction that uses audio, video, and an image.
const inputUri =
  'gs://cloud-samples-data/generative-ai/batch/batch_requests_for_multimodal_input.jsonl';
const location = 'us-central1';
const parent = `projects/${projectId}/locations/${location}`;
const modelName = `${parent}/publishers/google/models/gemini-1.5-flash-002`;

// Specify the location of the api endpoint.
const clientOptions = {
  apiEndpoint: `${location}-aiplatform.googleapis.com`,
};

// Instantiate the client.
const jobServiceClient = new aiplatformLib.JobServiceClient(clientOptions);

// Create a Gemini batch prediction job using Google Cloud Storage input and output buckets.
async function create_batch_prediction_gemini_gcs() {
  const gcsSource = new aiplatform.GcsSource({
    uris: [inputUri],
  });

  const inputConfig = new aiplatform.BatchPredictionJob.InputConfig({
    gcsSource: gcsSource,
    instancesFormat: 'jsonl',
  });

  const gcsDestination = new aiplatform.GcsDestination({
    outputUriPrefix: outputUri,
  });

  const outputConfig = new aiplatform.BatchPredictionJob.OutputConfig({
    gcsDestination: gcsDestination,
    predictionsFormat: 'jsonl',
  });

  const batchPredictionJob = new aiplatform.BatchPredictionJob({
    displayName: 'Batch predict with Gemini - GCS',
    model: modelName,
    inputConfig: inputConfig,
    outputConfig: outputConfig,
  });

  const request = {
    parent: parent,
    batchPredictionJob,
  };

  // Create batch prediction job request
  const [response] = await jobServiceClient.createBatchPredictionJob(request);
  console.log('Response name: ', response.name);
  // Example response:
  // Response name: projects/<project>/locations/us-central1/batchPredictionJobs/<job-id>
}

await create_batch_prediction_gemini_gcs();

Entrada do BigQuery

// Import the aiplatform library
const aiplatformLib = require('@google-cloud/aiplatform');
const aiplatform = aiplatformLib.protos.google.cloud.aiplatform.v1;

/**
 * TODO(developer):  Uncomment/update these variables before running the sample.
 */
// projectId = 'YOUR_PROJECT_ID';
// URI of the output BigQuery table.
// E.g. "bq://[PROJECT].[DATASET].[TABLE]"
// outputUri = 'bq://projectid.dataset.table';

// URI of the multimodal input BigQuery table.
// E.g. "bq://[PROJECT].[DATASET].[TABLE]"
const inputUri =
  'bq://storage-samples.generative_ai.batch_requests_for_multimodal_input';
const location = 'us-central1';
const parent = `projects/${projectId}/locations/${location}`;
const modelName = `${parent}/publishers/google/models/gemini-1.5-flash-002`;

// Specify the location of the api endpoint.
const clientOptions = {
  apiEndpoint: `${location}-aiplatform.googleapis.com`,
};

// Instantiate the client.
const jobServiceClient = new aiplatformLib.JobServiceClient(clientOptions);

// Create a Gemini batch prediction job using BigQuery input and output datasets.
async function create_batch_prediction_gemini_bq() {
  const bqSource = new aiplatform.BigQuerySource({
    inputUri: inputUri,
  });

  const inputConfig = new aiplatform.BatchPredictionJob.InputConfig({
    bigquerySource: bqSource,
    instancesFormat: 'bigquery',
  });

  const bqDestination = new aiplatform.BigQueryDestination({
    outputUri: outputUri,
  });

  const outputConfig = new aiplatform.BatchPredictionJob.OutputConfig({
    bigqueryDestination: bqDestination,
    predictionsFormat: 'bigquery',
  });

  const batchPredictionJob = new aiplatform.BatchPredictionJob({
    displayName: 'Batch predict with Gemini - BigQuery',
    model: modelName, // Add model parameters per request in the input BigQuery table.
    inputConfig: inputConfig,
    outputConfig: outputConfig,
  });

  const request = {
    parent: parent,
    batchPredictionJob,
  };

  // Create batch prediction job request
  const [response] = await jobServiceClient.createBatchPredictionJob(request);
  console.log('Response name: ', response.name);
  // Example response:
  // Response name: projects/<project>/locations/us-central1/batchPredictionJobs/<job-id>
}

await create_batch_prediction_gemini_bq();

Java

Antes de testar esse exemplo, siga as instruções de configuração para Java no Guia de início rápido da Vertex AI sobre como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Vertex AI para Java.

Para autenticar na Vertex AI, configure o Application Default Credentials. Para mais informações, consulte Configurar a autenticação para um ambiente de desenvolvimento local.

Entrada do Cloud Storage

import com.google.cloud.aiplatform.v1.BatchPredictionJob;
import com.google.cloud.aiplatform.v1.GcsDestination;
import com.google.cloud.aiplatform.v1.GcsSource;
import com.google.cloud.aiplatform.v1.JobServiceClient;
import com.google.cloud.aiplatform.v1.JobServiceSettings;
import com.google.cloud.aiplatform.v1.LocationName;
import java.io.IOException;

public class CreateBatchPredictionGeminiJobSample {

  public static void main(String[] args) throws IOException {
    // TODO(developer): Update these variables before running the sample.
    String project = "PROJECT_ID";
    String gcsDestinationOutputUriPrefix = "gs://MY_BUCKET/";

    createBatchPredictionGeminiJobSample(project, gcsDestinationOutputUriPrefix);
  }

  // Create a batch prediction job using a JSONL input file and output URI, both in Cloud
  // Storage.
  public static BatchPredictionJob createBatchPredictionGeminiJobSample(
      String project, String gcsDestinationOutputUriPrefix) throws IOException {
    String location = "us-central1";
    JobServiceSettings settings =
        JobServiceSettings.newBuilder()
            .setEndpoint(String.format("%s-aiplatform.googleapis.com:443", location))
            .build();

    // Initialize client that will be used to send requests. This client only needs to be created
    // once, and can be reused for multiple requests.
    try (JobServiceClient client = JobServiceClient.create(settings)) {
      GcsSource gcsSource =
          GcsSource.newBuilder()
              .addUris(
                  "gs://cloud-samples-data/generative-ai/batch/"
                      + "batch_requests_for_multimodal_input.jsonl")
              // Or try
              // "gs://cloud-samples-data/generative-ai/batch/gemini_multimodal_batch_predict.jsonl"
              // for a batch prediction that uses audio, video, and an image.
              .build();
      BatchPredictionJob.InputConfig inputConfig =
          BatchPredictionJob.InputConfig.newBuilder()
              .setInstancesFormat("jsonl")
              .setGcsSource(gcsSource)
              .build();
      GcsDestination gcsDestination =
          GcsDestination.newBuilder().setOutputUriPrefix(gcsDestinationOutputUriPrefix).build();
      BatchPredictionJob.OutputConfig outputConfig =
          BatchPredictionJob.OutputConfig.newBuilder()
              .setPredictionsFormat("jsonl")
              .setGcsDestination(gcsDestination)
              .build();
      String modelName =
          String.format(
              "projects/%s/locations/%s/publishers/google/models/%s",
              project, location, "gemini-1.5-flash-002");

      BatchPredictionJob batchPredictionJob =
          BatchPredictionJob.newBuilder()
              .setDisplayName("my-display-name")
              .setModel(modelName) // Add model parameters per request in the input jsonl file.
              .setInputConfig(inputConfig)
              .setOutputConfig(outputConfig)
              .build();

      LocationName parent = LocationName.of(project, location);
      BatchPredictionJob response = client.createBatchPredictionJob(parent, batchPredictionJob);
      System.out.format("\tName: %s\n", response.getName());
      // Example response:
      //   Name: projects/<project>/locations/us-central1/batchPredictionJobs/<job-id>
      return response;
    }
  }
}

Entrada do BigQuery

import com.google.cloud.aiplatform.v1.BatchPredictionJob;
import com.google.cloud.aiplatform.v1.BigQueryDestination;
import com.google.cloud.aiplatform.v1.BigQuerySource;
import com.google.cloud.aiplatform.v1.JobServiceClient;
import com.google.cloud.aiplatform.v1.JobServiceSettings;
import com.google.cloud.aiplatform.v1.LocationName;
import java.io.IOException;

public class CreateBatchPredictionGeminiBigqueryJobSample {

  public static void main(String[] args) throws IOException {
    // TODO(developer): Update these variables before running the sample.
    String project = "PROJECT_ID";
    String bigqueryDestinationOutputUri = "bq://PROJECT_ID.MY_DATASET.MY_TABLE";

    createBatchPredictionGeminiBigqueryJobSample(project, bigqueryDestinationOutputUri);
  }

  // Create a batch prediction job using BigQuery input and output datasets.
  public static BatchPredictionJob createBatchPredictionGeminiBigqueryJobSample(
      String project, String bigqueryDestinationOutputUri) throws IOException {
    String location = "us-central1";
    JobServiceSettings settings =
        JobServiceSettings.newBuilder()
            .setEndpoint(String.format("%s-aiplatform.googleapis.com:443", location))
            .build();

    // Initialize client that will be used to send requests. This client only needs to be created
    // once, and can be reused for multiple requests.
    try (JobServiceClient client = JobServiceClient.create(settings)) {
      BigQuerySource bigquerySource =
          BigQuerySource.newBuilder()
              .setInputUri("bq://storage-samples.generative_ai.batch_requests_for_multimodal_input")
              .build();
      BatchPredictionJob.InputConfig inputConfig =
          BatchPredictionJob.InputConfig.newBuilder()
              .setInstancesFormat("bigquery")
              .setBigquerySource(bigquerySource)
              .build();
      BigQueryDestination bigqueryDestination =
          BigQueryDestination.newBuilder().setOutputUri(bigqueryDestinationOutputUri).build();
      BatchPredictionJob.OutputConfig outputConfig =
          BatchPredictionJob.OutputConfig.newBuilder()
              .setPredictionsFormat("bigquery")
              .setBigqueryDestination(bigqueryDestination)
              .build();
      String modelName =
          String.format(
              "projects/%s/locations/%s/publishers/google/models/%s",
              project, location, "gemini-1.5-flash-002");

      BatchPredictionJob batchPredictionJob =
          BatchPredictionJob.newBuilder()
              .setDisplayName("my-display-name")
              .setModel(modelName) // Add model parameters per request in the input BigQuery table.
              .setInputConfig(inputConfig)
              .setOutputConfig(outputConfig)
              .build();

      LocationName parent = LocationName.of(project, location);
      BatchPredictionJob response = client.createBatchPredictionJob(parent, batchPredictionJob);
      System.out.format("\tName: %s\n", response.getName());
      // Example response:
      //   Name: projects/<project>/locations/us-central1/batchPredictionJobs/<job-id>
      return response;
    }
  }
}

Go

Antes de testar esse exemplo, siga as instruções de configuração para Go no Guia de início rápido da Vertex AI sobre como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Vertex AI para Go.

Para autenticar na Vertex AI, configure o Application Default Credentials. Para mais informações, consulte Configurar a autenticação para um ambiente de desenvolvimento local.

Entrada do Cloud Storage

import (
	"context"
	"fmt"
	"io"
	"time"

	aiplatform "cloud.google.com/go/aiplatform/apiv1"
	aiplatformpb "cloud.google.com/go/aiplatform/apiv1/aiplatformpb"

	"google.golang.org/api/option"
	"google.golang.org/protobuf/types/known/structpb"
)

// batchPredictGCS submits a batch prediction job using GCS data source as its input
func batchPredictGCS(w io.Writer, projectID, location string, inputURIs []string, outputURI string) error {
	// location := "us-central1"
	// inputURIs := []string{"gs://cloud-samples-data/batch/prompt_for_batch_gemini_predict.jsonl"}
	// outputURI := "gs://<cloud-bucket-name>/<prefix-name>"
	modelName := "gemini-1.5-pro-002"
	jobName := "batch-predict-gcs-test-001"

	ctx := context.Background()
	apiEndpoint := fmt.Sprintf("%s-aiplatform.googleapis.com:443", location)
	client, err := aiplatform.NewJobClient(ctx, option.WithEndpoint(apiEndpoint))
	if err != nil {
		return fmt.Errorf("unable to create aiplatform client: %w", err)
	}
	defer client.Close()

	modelParameters, err := structpb.NewValue(map[string]interface{}{
		"temperature":     0.2,
		"maxOutputTokens": 200,
	})
	if err != nil {
		return fmt.Errorf("unable to convert model parameters to protobuf value: %w", err)
	}

	req := &aiplatformpb.CreateBatchPredictionJobRequest{
		Parent: fmt.Sprintf("projects/%s/locations/%s", projectID, location),
		BatchPredictionJob: &aiplatformpb.BatchPredictionJob{
			DisplayName:     jobName,
			Model:           fmt.Sprintf("publishers/google/models/%s", modelName),
			ModelParameters: modelParameters,
			// Check the API reference for `BatchPredictionJob` for supported input and output formats:
			// https://cloud.google.com/vertex-ai/docs/reference/rpc/google.cloud.aiplatform.v1#google.cloud.aiplatform.v1.BatchPredictionJob
			InputConfig: &aiplatformpb.BatchPredictionJob_InputConfig{
				Source: &aiplatformpb.BatchPredictionJob_InputConfig_GcsSource{
					GcsSource: &aiplatformpb.GcsSource{
						Uris: inputURIs,
					},
				},
				InstancesFormat: "jsonl",
			},
			OutputConfig: &aiplatformpb.BatchPredictionJob_OutputConfig{
				Destination: &aiplatformpb.BatchPredictionJob_OutputConfig_GcsDestination{
					GcsDestination: &aiplatformpb.GcsDestination{
						OutputUriPrefix: outputURI,
					},
				},
				PredictionsFormat: "jsonl",
			},
		},
	}

	job, err := client.CreateBatchPredictionJob(ctx, req)
	if err != nil {
		return err
	}
	fullJobId := job.GetName()
	fmt.Fprintf(w, "submitted batch predict job for model %q\n", job.GetModel())
	fmt.Fprintf(w, "job id: %q\n", fullJobId)
	fmt.Fprintf(w, "job state: %s\n", job.GetState())
	// Example response:
	// submitted batch predict job for model "publishers/google/models/gemini-1.5-pro-002"
	// job id: "projects/.../locations/.../batchPredictionJobs/1234567890000000000"
	// job state: JOB_STATE_PENDING

	for {
		time.Sleep(5 * time.Second)

		job, err := client.GetBatchPredictionJob(ctx, &aiplatformpb.GetBatchPredictionJobRequest{
			Name: fullJobId,
		})
		if err != nil {
			return fmt.Errorf("error: couldn't get updated job state: %w", err)
		}

		if job.GetEndTime() != nil {
			fmt.Fprintf(w, "batch predict job finished with state %s\n", job.GetState())
			break
		} else {
			fmt.Fprintf(w, "batch predict job is running... job state is %s\n", job.GetState())
		}
	}

	return nil
}

Entrada do BigQuery

import (
	"context"
	"fmt"
	"io"
	"time"

	aiplatform "cloud.google.com/go/aiplatform/apiv1"
	aiplatformpb "cloud.google.com/go/aiplatform/apiv1/aiplatformpb"

	"google.golang.org/api/option"
	"google.golang.org/protobuf/types/known/structpb"
)

// batchPredictBQ submits a batch prediction job using BigQuery data source as its input
func batchPredictBQ(w io.Writer, projectID, location string, inputURI string, outputURI string) error {
	// location  := "us-central1"
	// inputURI  := "bq://storage-samples.generative_ai.batch_requests_for_multimodal_input"
	// outputURI := "bq://<cloud-project-name>.<dataset-name>.<table-name>"
	modelName := "gemini-1.5-pro-002"
	jobName := "batch-predict-bq-test-001"

	ctx := context.Background()
	apiEndpoint := fmt.Sprintf("%s-aiplatform.googleapis.com:443", location)
	client, err := aiplatform.NewJobClient(ctx, option.WithEndpoint(apiEndpoint))
	if err != nil {
		return fmt.Errorf("unable to create aiplatform client: %w", err)
	}
	defer client.Close()

	modelParameters, err := structpb.NewValue(map[string]interface{}{
		"temperature":     0.2,
		"maxOutputTokens": 200,
	})
	if err != nil {
		return fmt.Errorf("unable to convert model parameters to protobuf value: %w", err)
	}

	req := &aiplatformpb.CreateBatchPredictionJobRequest{
		Parent: fmt.Sprintf("projects/%s/locations/%s", projectID, location),
		BatchPredictionJob: &aiplatformpb.BatchPredictionJob{
			DisplayName:     jobName,
			Model:           fmt.Sprintf("publishers/google/models/%s", modelName),
			ModelParameters: modelParameters,
			// Check the API reference for `BatchPredictionJob` for supported input and output formats:
			// https://cloud.google.com/vertex-ai/docs/reference/rpc/google.cloud.aiplatform.v1#google.cloud.aiplatform.v1.BatchPredictionJob
			InputConfig: &aiplatformpb.BatchPredictionJob_InputConfig{
				Source: &aiplatformpb.BatchPredictionJob_InputConfig_BigquerySource{
					BigquerySource: &aiplatformpb.BigQuerySource{
						InputUri: inputURI,
					},
				},
				InstancesFormat: "bigquery",
			},

			OutputConfig: &aiplatformpb.BatchPredictionJob_OutputConfig{
				Destination: &aiplatformpb.BatchPredictionJob_OutputConfig_BigqueryDestination{
					BigqueryDestination: &aiplatformpb.BigQueryDestination{
						OutputUri: outputURI,
					},
				},
				PredictionsFormat: "bigquery",
			},
		},
	}

	job, err := client.CreateBatchPredictionJob(ctx, req)
	if err != nil {
		return err
	}
	fullJobId := job.GetName()
	fmt.Fprintf(w, "submitted batch predict job for model %q\n", job.GetModel())
	fmt.Fprintf(w, "job id: %q\n", fullJobId)
	fmt.Fprintf(w, "job state: %s\n", job.GetState())
	// Example response:
	// submitted batch predict job for model "publishers/google/models/gemini-1.5-pro-002"
	// job id: "projects/.../locations/.../batchPredictionJobs/1234567890000000000"
	// job state: JOB_STATE_PENDING

	for {
		time.Sleep(5 * time.Second)

		job, err := client.GetBatchPredictionJob(ctx, &aiplatformpb.GetBatchPredictionJobRequest{
			Name: fullJobId,
		})
		if err != nil {
			return fmt.Errorf("error: couldn't get updated job state: %w", err)
		}

		if job.GetEndTime() != nil {
			fmt.Fprintf(w, "batch predict job finished with state %s\n", job.GetState())
			break
		} else {
			fmt.Fprintf(w, "batch predict job is running... job state is %s\n", job.GetState())
		}
	}

	return nil
}

Recuperar saída em lote

Quando uma tarefa de previsão em lote é concluída, a saída é armazenada no bucket do Cloud Storage ou na tabela do BigQuery especificada na solicitação.

A seguir