How can I pass request body for sklearn model, decision tree classifier in AWS Sagemaker while invoking endpoint?

Asked Jun 01 '22 at 10:29

Active Jun 01 '22 at 10:29

Viewed 155 times

I am trying to invoke a multi-model endpoint in AWS Sagemaker and I am getting a ModelError -ModelError: An error occurred (ModelError) when calling the InvokeEndpoint operation: Received server error (500) from model with message "invalid literal for int() with base 10: 'Iris-virginica'". I have used Sklearn model with IRIS dataset and Decision Tree Algorithm to predict the output.

import boto3
import json

runtime_client = boto3.client('sagemaker')
content_type = "application/json"
request_body = {"Input": [[1.2,0.5, 1.7, 2.6, 2.7,1.0]]}
data = json.loads(json.dumps(request_body))
payload = json.dumps(data)
endpoint_name = "sklearn-endpoint"
print("payload ",type(payload))
response = runtime_client.invoke_endpoint(
    EndpointName=endpoint_name,
    ContentType=content_type,
    TargetModel="iris.tar.gz",
    Body=payload)

asked Jun 01 '22 at 10:29

Mrunall Veer

Can you try putting the floats in your vector in quotes so they are passed to the endpoint as strings? – Kirit Thadaka Jun 01 '22 at 18:55
Can you please elaborate may be with an example? – Mrunall Veer Jun 02 '22 at 08:42
Try changing - `request_body = {"Input": [[1.2,0.5, 1.7, 2.6, 2.7,1.0]]}` to `request_body = {"Input": [['1.2','0.5', '1.7', '2.6', '2.7','1.0']]}` – Kirit Thadaka Jun 07 '22 at 21:50

How can I pass request body for sklearn model, decision tree classifier in AWS Sagemaker while invoking endpoint?

0 Answers0