Unable to deploy Document Understanding ML package in AI Center

Hello there,
I am trying to create an ML Skill with the Document Understanding ML package available under Out-Of-The-Box packages in AI Center. But, the deployment is failing with the below error in ML logs.

It would be great if someone can help me in resolving this.

“Failed to create deployment DocumentUnderstanding”
"File “/microservice/main.py”, line 14, in init
service_internal.init(opt_in=opt, args=argv.ArgvParser().parse(args))
File “/microservice/extraction/model_tag/service_internal.py”, line 120, in init
raise ServiceException(error_msg, error_code)
util.misc.ServiceException: Service Exception 500: ModelLoadFailure - Unable to load extraction model for: default, ensure that the specified model has been trained

[2021-02-04 08:27:42 +0000] [22] [ERROR] Exception in worker process
File “/microservice/main.py”, line 14, in init
service_internal.init(opt_in=opt, args=argv.ArgvParser().parse(args))
File “/microservice/extraction/model_tag/service_internal.py”, line 120, in init
raise ServiceException(error_msg, error_code)
util.misc.ServiceException: Service Exception 500: ModelLoadFailure - Unable to load extraction model for: default, ensure that the specified model has been trained
[2021-02-04 08:27:42 +0000] [22] [INFO] Worker exiting (pid: 22)"

Hi @bheema

This model needs to be trained before you can deploy it, if you don’t have a dataset available I would suggest that you do your testing with another Document Understanding model (e.g.: Invoices, Receipts…) that can be deployed right away.

Jeremy

Thanks so much for providing the details for the error.

Actually my requirement is to extract the details from the checks using AI Fabric. So, I can’t use Invoice/Receipt packages.

It would be great if u suggest which ML package is suitable for this. Or provide sample dataset for training the document understanding package.

Thanks in advance.

Hi @bheema,
do you find any solution to your custom ML model? Because I am stuck in this same situation (extract different information using DocumentUnderstanding package)