SageMaker：调用 InvokeEndpoint 操作时发生错误 (ModelError)：无法评估提供的负载

Question

我在 Jupyter notebook 的 Amazon SageMaker（图像分类算法）中有一个运行良好的端点。在 Lambda 函数中也工作正常，当我从 API 网关调用 Lambda 函数时，从 API 网关的测试中，也工作正常。

问题是当我根据这个答案从 Postman 调用 API 时：

Lambda 中的代码是：

import boto3
import json
import base64

ENDPOINT_NAME = "DEMO-XGBoostEndpoint-Multilabel"
runtime= boto3.client("runtime.sagemaker")
imagen_ = "/tmp/imageToProcess.jpg"

def write_to_file(save_path, data):
    with open(save_path, "wb") as f:
        f.write(base64.b64decode(data))

def lambda_handler(event, context):
    img_json = json.loads(json.dumps(event))

    write_to_file(imagen_, json.dumps(event, indent=2))

    with open(imagen_, "rb") as image:
        f = image.read()
        b = bytearray(f)

    payload = b

    response = runtime.invoke_endpoint(EndpointName=ENDPOINT_NAME,
                                       ContentType="application/x-image",
                                       Body=payload)

    #print(response)
    result = json.loads(response["Body"].read().decode())
    print(result)
    predicted_label=[]
    classes = ["chair", "handbag", "person", "traffic light", "clock"]
    for idx, val in enumerate(classes):
        print("%s:%f "%(classes[idx], result[idx]), end="")
        predicted_label += (classes[idx], result[idx])

    return {
      "statusCode": 200,
      "headers": { "content-type": "application/json"},
      "body":  predicted_label
}

错误是：

Traceback (most recent call last):
  File "/var/task/lambda_function.py", line 26, in lambda_handler
    Body=payload)
  File "/var/runtime/botocore/client.py", line 316, in _api_call
    return self._make_api_call(operation_name, kwargs)
  File "/var/runtime/botocore/client.py", line 626, in _make_api_call
    raise error_class(parsed_response, operation_name)
botocore.errorfactory.ModelError: An error occurred (ModelError) when calling the InvokeEndpoint operation: Received client error (400) from model with message "unable to evaluate payload provided". See https://us-east-2.console.aws.amazon.com/cloudwatch/home?region=us-east-2#logEventViewer:group=/aws/sagemaker/Endpoints/DEMO-XGBoostEndpoint-Multilabel in account 866341179300 for more information. ```

Answer 1

我用 this post:

解决了

谢谢大家

最后lambda函数中的代码是：

import os
import boto3
import json
import base64

ENDPOINT_NAME = os.environ['endPointName']
CLASSES = "["chair", "handbag", "person", "traffic light", "clock"]"
runtime= boto3.client("runtime.sagemaker")

def lambda_handler(event, context):
    file_content = base64.b64decode(event['content'])

    payload = file_content
    response = runtime.invoke_endpoint(EndpointName=ENDPOINT_NAME, ContentType="application/x-image", Body=payload)

    result = json.loads(response["Body"].read().decode())
    print(result)
    predicted_label=[]
    classes = CLASSES
    for idx, val in enumerate(classes):
       print("%s:%f "%(classes[idx], result[idx]), end="")
       predicted_label += (classes[idx], result[idx])

    return {
      "statusCode": 200,
      "headers": { "content-type": "application/json"},
      "body":  predicted_label
    }

SageMaker：调用 InvokeEndpoint 操作时发生错误 (ModelError)：无法评估提供的负载

SageMaker: An error occurred (ModelError) when calling the InvokeEndpoint operation: unable to evaluate payload provided

amazon-web-services

aws-lambda

amazon-sagemaker