Training failed for pipeline with Error: 'multi_task Extraction Model Training Failed'

Help AI Center
, , ,
1

Hi all,

I’m creating an ML Skill for processing passports. There are 4 FOIs to be extracted from a passport and the document format is PDF. For this I’m using an OOB Package for Passport (version: 22.10.1).
After executing the Train Only pipeline, I’m getting an error as mentioned in the screenshot below:

image
image949×419 57.3 KB

For data labelling, I have used the Document Manager and exported the labelled schema into AI Center. (refer dataset screenshot below)

image
image947×412 14.2 KB

Please help me with this!

Thanks.

2

@ajay_anil

Assertion error generally comes when the given dataset is wrong…Can you check the dataset given is proper and the uploaded schema is correct

cheers

3

@Anil_G Thanks for the response.

Is there any specific file/content that I need to check within the dataset exported .
I have re-uploaded the dataset couple of times and also tried different options in the document manager while exporting.(schema and all labelled)

Please assist.

4

Hi @ajay_anil did you figure out what caused this issue?