aws-samples / amazon-rekognition-custom-labels-feedback-solution Goto Github PK

The documentation says "Deploy CloudFormation stack in one of the AWS regions where you are using Amazon Rekognition Custom Labels." But the deployment fails if i change the region in the link to eu-west-1.

It fails on the step GTLabelVerificationPreLambda with error: Error occurred while GetObject. S3 Error Code: PermanentRedirect. S3 Error Message: The bucket is in this region: us-east-1. Please use this region to retry the request (Service: AWSLambdaInternal; Status Code: 400; Error Code: InvalidParameterValueException; Request ID: 2c779ae6-b8e2-4462-8721-a0bc13ee1e5a)

Hi, I am trying to use the manifest output of this code to gather label, bounding box, and confidence data for a large number of images.
Why do the manifest files all show a fixed confidence of 0.9 rather than that actual label confidence?
Line 422 in 'start-feedback.py' just appends 0.9 rather than a dynamic confidence from the model.

This code would be really helpful if there was some additional documentation for people who are not strong coders, but needing to analyse images using ML. Any suggestions or pointers would be very welcome and would help with a conservation project.

The README isn't clear enough. In particular, I don't know what value should go to images and outputBucket fields on feedback-config.json