Invoke Model API를 사용하여 Amazon Bedrock에서 Mistral AI 모델 간접 호출

다음 코드 예제에서는 Invoke Model API를 사용하여 Mistral 모델에 텍스트 메시지를 보내는 방법을 보여줍니다.

.NET

AWS SDK for .NET

참고

GitHub에 더 많은 내용이 있습니다. AWS 코드 예시 리포지토리에서 전체 예시를 찾고 설정 및 실행하는 방법을 배워보세요.

Invoke Model API를 사용하여 텍스트 메시지를 보냅니다.

// Use the native inference API to send a text message to Mistral.

using System;
using System.IO;
using System.Text.Json;
using System.Text.Json.Nodes;
using Amazon;
using Amazon.BedrockRuntime;
using Amazon.BedrockRuntime.Model;

// Create a Bedrock Runtime client in the AWS Region you want to use.
var client = new AmazonBedrockRuntimeClient(RegionEndpoint.USEast1);

// Set the model ID, e.g., Mistral Large.
var modelId = "mistral.mistral-large-2402-v1:0";

// Define the prompt for the model.
var prompt = "Describe the purpose of a 'hello world' program in one line.";

// Embed the prompt in Mistral's instruction format.
var formattedPrompt = $"<s>[INST] {prompt} [/INST]";

//Format the request payload using the model's native structure.
var nativeRequest = JsonSerializer.Serialize(new
{
    prompt = formattedPrompt,
    max_tokens = 512,
    temperature = 0.5
});

// Create a request with the model ID and the model's native request payload.
var request = new InvokeModelRequest()
{
    ModelId = modelId,
    Body = new MemoryStream(System.Text.Encoding.UTF8.GetBytes(nativeRequest)),
    ContentType = "application/json"
};

try
{
    // Send the request to the Bedrock Runtime and wait for the response.
    var response = await client.InvokeModelAsync(request);

    // Decode the response body.
    var modelResponse = await JsonNode.ParseAsync(response.Body);

    // Extract and print the response text.
    var responseText = modelResponse["outputs"]?[0]?["text"] ?? "";
    Console.WriteLine(responseText);
}
catch (AmazonBedrockRuntimeException e)
{
    Console.WriteLine($"ERROR: Can't invoke '{modelId}'. Reason: {e.Message}");
    throw;
}

• API 세부 정보는 AWS SDK for .NET API 참조의 InvokeModel을 참조하세요.

Java

SDK for Java 2.x

참고

GitHub에 더 많은 내용이 있습니다. AWS 코드 예시 리포지토리에서 전체 예시를 찾고 설정 및 실행하는 방법을 배워보세요.

Invoke Model API를 사용하여 텍스트 메시지를 보냅니다.

// Use the native inference API to send a text message to Mistral.

import org.json.JSONObject;
import org.json.JSONPointer;
import software.amazon.awssdk.auth.credentials.DefaultCredentialsProvider;
import software.amazon.awssdk.core.SdkBytes;
import software.amazon.awssdk.core.exception.SdkClientException;
import software.amazon.awssdk.regions.Region;
import software.amazon.awssdk.services.bedrockruntime.BedrockRuntimeClient;

public class InvokeModel {

    public static String invokeModel() {

        // Create a Bedrock Runtime client in the AWS Region you want to use.
        // Replace the DefaultCredentialsProvider with your preferred credentials provider.
        var client = BedrockRuntimeClient.builder()
                .credentialsProvider(DefaultCredentialsProvider.create())
                .region(Region.US_EAST_1)
                .build();

        // Set the model ID, e.g., Mistral Large.
        var modelId = "mistral.mistral-large-2402-v1:0";

        // The InvokeModel API uses the model's native payload.
        // Learn more about the available inference parameters and response fields at:
        // https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-mistral-text-completion.html
        var nativeRequestTemplate = "{ \"prompt\": \"{{instruction}}\" }";

        // Define the prompt for the model.
        var prompt = "Describe the purpose of a 'hello world' program in one line.";

        // Embed the prompt in Mistral's instruction format.
        var instruction = "<s>[INST] {{prompt}} [/INST]\\n".replace("{{prompt}}", prompt);

        // Embed the instruction in the the native request payload.
        var nativeRequest = nativeRequestTemplate.replace("{{instruction}}", instruction);

        try {
            // Encode and send the request to the Bedrock Runtime.
            var response = client.invokeModel(request -> request
                    .body(SdkBytes.fromUtf8String(nativeRequest))
                    .modelId(modelId)
            );

            // Decode the response body.
            var responseBody = new JSONObject(response.body().asUtf8String());

            // Retrieve the generated text from the model's response.
            var text = new JSONPointer("/outputs/0/text").queryFrom(responseBody).toString();
            System.out.println(text);

            return text;

        } catch (SdkClientException e) {
            System.err.printf("ERROR: Can't invoke '%s'. Reason: %s", modelId, e.getMessage());
            throw new RuntimeException(e);
        }
    }

    public static void main(String[] args) {
        invokeModel();
    }
}

• API 세부 정보는 AWS SDK for Java 2.x API 참조의 InvokeModel을 참조하세요.

JavaScript

SDK for JavaScript (v3)

참고

GitHub에 더 많은 내용이 있습니다. AWS 코드 예시 리포지토리에서 전체 예시를 찾고 설정 및 실행하는 방법을 배워보세요.

Invoke Model API를 사용하여 텍스트 메시지를 보냅니다.

import { fileURLToPath } from "node:url";

import { FoundationModels } from "../../config/foundation_models.js";
import {
  BedrockRuntimeClient,
  InvokeModelCommand,
} from "@aws-sdk/client-bedrock-runtime";

/**
 * @typedef {Object} Output
 * @property {string} text
 *
 * @typedef {Object} ResponseBody
 * @property {Output[]} outputs
 */

/**
 * Invokes a Mistral 7B Instruct model.
 *
 * @param {string} prompt - The input text prompt for the model to complete.
 * @param {string} [modelId] - The ID of the model to use. Defaults to "mistral.mistral-7b-instruct-v0:2".
 */
export const invokeModel = async (
  prompt,
  modelId = "mistral.mistral-7b-instruct-v0:2",
) => {
  // Create a new Bedrock Runtime client instance.
  const client = new BedrockRuntimeClient({ region: "us-east-1" });

  // Mistral instruct models provide optimal results when embedding
  // the prompt into the following template:
  const instruction = `<s>[INST] ${prompt} [/INST]`;

  // Prepare the payload.
  const payload = {
    prompt: instruction,
    max_tokens: 500,
    temperature: 0.5,
  };

  // Invoke the model with the payload and wait for the response.
  const command = new InvokeModelCommand({
    contentType: "application/json",
    body: JSON.stringify(payload),
    modelId,
  });
  const apiResponse = await client.send(command);

  // Decode and return the response.
  const decodedResponseBody = new TextDecoder().decode(apiResponse.body);
  /** @type {ResponseBody} */
  const responseBody = JSON.parse(decodedResponseBody);
  return responseBody.outputs[0].text;
};

// Invoke the function if this file was run directly.
if (process.argv[1] === fileURLToPath(import.meta.url)) {
  const prompt =
    'Complete the following in one sentence: "Once upon a time..."';
  const modelId = FoundationModels.MISTRAL_7B.modelId;
  console.log(`Prompt: ${prompt}`);
  console.log(`Model ID: ${modelId}`);

  try {
    console.log("-".repeat(53));
    const response = await invokeModel(prompt, modelId);
    console.log(response);
  } catch (err) {
    console.log(err);
  }
}

• API 세부 정보는 AWS SDK for JavaScript API 참조의 InvokeModel을 참조하세요.

Python

SDK for Python (Boto3)

참고

GitHub에 더 많은 내용이 있습니다. AWS 코드 예시 리포지토리에서 전체 예시를 찾고 설정 및 실행하는 방법을 배워보세요.

Invoke Model API를 사용하여 텍스트 메시지를 보냅니다.

# Use the native inference API to send a text message to Mistral.

import boto3
import json
from botocore.exceptions import ClientError

# Create a Bedrock Runtime client in the AWS Region of your choice.
client = boto3.client("bedrock-runtime", region_name="us-east-1")

# Set the model ID, e.g., Mistral Large.
model_id = "mistral.mistral-large-2402-v1:0"

# Define the prompt for the model.
prompt = "Describe the purpose of a 'hello world' program in one line."

# Embed the prompt in Mistral's instruction format.
formatted_prompt = f"<s>[INST] {prompt} [/INST]"

# Format the request payload using the model's native structure.
native_request = {
    "prompt": formatted_prompt,
    "max_tokens": 512,
    "temperature": 0.5,
}

# Convert the native request to JSON.
request = json.dumps(native_request)

try:
    # Invoke the model with the request.
    response = client.invoke_model(modelId=model_id, body=request)

except (ClientError, Exception) as e:
    print(f"ERROR: Can't invoke '{model_id}'. Reason: {e}")
    exit(1)

# Decode the response body.
model_response = json.loads(response["body"].read())

# Extract and print the response text.
response_text = model_response["outputs"][0]["text"]
print(response_text)

• API 세부 정보는 AWS SDK for Python (Boto3) API 참조의 InvokeModel를 참조하세요.

AWS SDK 개발자 안내서 및 코드 예제의 전체 목록은 섹션을 참조하세요AWS SDK에서 Amazon Bedrock 사용. 이 주제에는 시작하기에 대한 정보와 이전 SDK 버전에 대한 세부 정보도 포함되어 있습니다.