Unable to deploy ML Skill for Arabic Documents

[ahmadiqbal.zuberi (Ahmad Zuberi)]

Hi, I am making ML Skill for extracting data from Arabic Documents. The labelling part was done. Pipeline ran successful but ML Skill is failing. I am using Out of the Box ML Package (Document Understanding). Is there any other specific ML package for Arabic Documents?

[supermanPunch (Arpan)]

@ahmadiqbal.zuberi

Pipeline ran successful but ML Skill is failing. I am using Out of the Box ML Package (Document Understanding). Is there any other specific ML package for Arabic Documents?

Hi @ahmadiqbal.zuberi ,

Could you let us know what was the Error Message received ?

[ahmadiqbal.zuberi (Ahmad Zuberi)]

Hi @supermanPunch,

Below is the screenshot of the error.

[supermanPunch (Arpan)]

@ahmadiqbal.zuberi ,

There are some posts here which mentions the same Issue, Did you check it out already ? :

AI Center - ML Skill Failed / Kubernetes operation failed to create deployment

Hi, We do not select the option (Make this an Evaluation Set) while importing your documents if you want to run your pipeline as Train run. [image] Make Sure that, the pipeline will be run based on the Data set Import. If you Choose as Evaluation set, The pipeline will get success only if you choose Evaluation run. The Suggested way is - to use Train run to improve your ML Skill and the evaluation run does not improve your ML Skill. Error Occurred : Resolved ML Skill Failed / Kubernete…

[ahmadiqbal.zuberi (Ahmad Zuberi)]

Yes @supermanPunch, I followed the same instructions.
When I am creating an ML Skill for English documents, it’s working as expected. However, on Arabic documents, it’s failing.