일괄 처리를 사용하여 텍스트에서 임베딩 생성

이 코드 샘플은 사전 학습된 모델을 사용하여 텍스트 입력 목록의 임베딩을 일괄 생성하고 지정된 위치에 저장하는 방법을 보여줍니다.

더 살펴보기

이 코드 샘플이 포함된 자세한 문서는 다음을 참조하세요.

코드 샘플

Java

이 샘플을 사용해 보기 전에 Vertex AI 빠른 시작: 클라이언트 라이브러리 사용Java 설정 안내를 따르세요. 자세한 내용은 Vertex AI Java API 참고 문서를 참조하세요.

Vertex AI에 인증하려면 애플리케이션 기본 사용자 인증 정보를 설정합니다. 자세한 내용은 로컬 개발 환경의 인증 설정을 참조하세요.


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 EmbeddingBatchSample {

  public static void main(String[] args) throws IOException, InterruptedException {
    // TODO(developer): Replace these variables before running the sample.
    String project = "YOUR_PROJECT_ID";
    String location = "us-central1";
    // inputUri: URI of the input dataset.
    // Could be a BigQuery table or a Google Cloud Storage file.
    // E.g. "gs://[BUCKET]/[DATASET].jsonl" OR "bq://[PROJECT].[DATASET].[TABLE]"
    String inputUri = "gs://cloud-samples-data/generative-ai/embeddings/embeddings_input.jsonl";
    // outputUri: URI where the output will be stored.
    // Could be a BigQuery table or a Google Cloud Storage file.
    // E.g. "gs://[BUCKET]/[OUTPUT].jsonl" OR "bq://[PROJECT].[DATASET].[TABLE]"
    String outputUri = "gs://YOUR_BUCKET/embedding_batch_output";
    String textEmbeddingModel = "text-embedding-005";

    embeddingBatchSample(project, location, inputUri, outputUri, textEmbeddingModel);
  }

  // Generates embeddings from text using batch processing.
  // Read more: https://cloud.google.com/vertex-ai/generative-ai/docs/embeddings/batch-prediction-genai-embeddings
  public static BatchPredictionJob embeddingBatchSample(
      String project, String location, String inputUri, String outputUri, String textEmbeddingModel)
      throws IOException {
    BatchPredictionJob response;
    JobServiceSettings jobServiceSettings =  JobServiceSettings.newBuilder()
        .setEndpoint("us-central1-aiplatform.googleapis.com:443").build();
    LocationName parent = LocationName.of(project, location);
    String modelName = String.format("projects/%s/locations/%s/publishers/google/models/%s",
        project, location, textEmbeddingModel);

    // 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(jobServiceSettings)) {
      BatchPredictionJob batchPredictionJob =
          BatchPredictionJob.newBuilder()
              .setDisplayName("my embedding batch job " + System.currentTimeMillis())
              .setModel(modelName)
              .setInputConfig(
                  BatchPredictionJob.InputConfig.newBuilder()
                      .setGcsSource(GcsSource.newBuilder().addUris(inputUri).build())
                      .setInstancesFormat("jsonl")
                      .build())
              .setOutputConfig(
                  BatchPredictionJob.OutputConfig.newBuilder()
                      .setGcsDestination(GcsDestination.newBuilder()
                          .setOutputUriPrefix(outputUri).build())
                      .setPredictionsFormat("jsonl")
                      .build())
              .build();

      response = client.createBatchPredictionJob(parent, batchPredictionJob);

      System.out.format("response: %s\n", response);
      System.out.format("\tName: %s\n", response.getName());
    }
    return response;
  }
}

Node.js

이 샘플을 사용해 보기 전에 Vertex AI 빠른 시작: 클라이언트 라이브러리 사용Node.js 설정 안내를 따르세요. 자세한 내용은 Vertex AI Node.js API 참고 문서를 참조하세요.

Vertex AI에 인증하려면 애플리케이션 기본 사용자 인증 정보를 설정합니다. 자세한 내용은 로컬 개발 환경의 인증 설정을 참조하세요.

// Imports 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';

// Optional: URI of the input dataset.
// Could be a BigQuery table or a Google Cloud Storage file.
// E.g. "gs://[BUCKET]/[DATASET].jsonl" OR "bq://[PROJECT].[DATASET].[TABLE]"
// inputUri =
//   'gs://cloud-samples-data/generative-ai/embeddings/embeddings_input.jsonl';

// Optional: URI where the output will be stored.
// Could be a BigQuery table or a Google Cloud Storage file.
// E.g. "gs://[BUCKET]/[OUTPUT].jsonl" OR "bq://[PROJECT].[DATASET].[TABLE]"
// outputUri = 'gs://your_bucket/embedding_batch_output';

// The name of the job
// jobName = `Batch embedding job: ${new Date().getMilliseconds()}`;

const textEmbeddingModel = 'text-embedding-005';
const location = 'us-central1';

// Configure the parent resource
const parent = `projects/${projectId}/locations/${location}`;
const modelName = `projects/${projectId}/locations/${location}/publishers/google/models/${textEmbeddingModel}`;

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

// Instantiates a client
const jobServiceClient = new aiplatformLib.JobServiceClient(clientOptions);

// Generates embeddings from text using batch processing.
// Read more: https://cloud.google.com/vertex-ai/generative-ai/docs/embeddings/batch-prediction-genai-embeddings
async function callBatchEmbedding() {
  const gcsSource = new aiplatform.GcsSource({
    uris: [inputUri],
  });

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

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

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

  const batchPredictionJob = new aiplatform.BatchPredictionJob({
    displayName: jobName,
    model: modelName,
    inputConfig,
    outputConfig,
  });

  const request = {
    parent,
    batchPredictionJob,
  };

  // Create batch prediction job request
  const [response] = await jobServiceClient.createBatchPredictionJob(request);

  console.log('Raw response: ', JSON.stringify(response, null, 2));
}

await callBatchEmbedding();

Python

이 샘플을 사용해 보기 전에 Vertex AI 빠른 시작: 클라이언트 라이브러리 사용Python 설정 안내를 따르세요. 자세한 내용은 Vertex AI Python API 참고 문서를 참조하세요.

Vertex AI에 인증하려면 애플리케이션 기본 사용자 인증 정보를 설정합니다. 자세한 내용은 로컬 개발 환경의 인증 설정을 참조하세요.

import vertexai

from vertexai.preview import language_models

# TODO(developer): Update & uncomment line below
# PROJECT_ID = "your-project-id"
vertexai.init(project=PROJECT_ID, location="us-central1")
input_uri = (
    "gs://cloud-samples-data/generative-ai/embeddings/embeddings_input.jsonl"
)
# Format: `"gs://your-bucket-unique-name/directory/` or `bq://project_name.llm_dataset`
output_uri = OUTPUT_URI

textembedding_model = language_models.TextEmbeddingModel.from_pretrained(
    "textembedding-gecko@003"
)

batch_prediction_job = textembedding_model.batch_predict(
    dataset=[input_uri],
    destination_uri_prefix=output_uri,
)
print(batch_prediction_job.display_name)
print(batch_prediction_job.resource_name)
print(batch_prediction_job.state)
# Example response:
# BatchPredictionJob 2024-09-10 15:47:51.336391
# projects/1234567890/locations/us-central1/batchPredictionJobs/123456789012345
# JobState.JOB_STATE_SUCCEEDED

다음 단계

다른 Google Cloud 제품의 코드 샘플을 검색하고 필터링하려면 Google Cloud 샘플 브라우저를 참조하세요.