AI Center x Document Understanding

[Pranav_Kende]

Hi All,

Steps that I took in Document Understanding:

1. Created a document type and labelled it.
2. Exported the dataset to AI Center
3. In Extractor tab, created a new Extractor on the document type with “Automated Training”

In AI Center:

1. Due to automated training, one training pipeline was success

After this, I have uploaded a PDF documents folder in the dataset. I am using this dataset for evaluation run. But if I am wanting to do an evaluation run, the pipeline is getting failed. with the error messages in the logs as below :

2024-03-27 15:51:03,583 - root:_valid_doctype_folder_structure:101 - ERROR: schema.json is empty / does not exist for bills_of_lading dataset

2024-03-27 15:51:03,583 - UiPath_core.training_plugin:model_run:189 - ERROR: Training failed for pipeline type: EVALUATE_ONLY, error: Document type bills_of_lading not valid, check that document type data is in dataset folder and follows folder structure

2024-03-27 15:51:03,584 - UiPath_core.trainer_run:main:100 - ERROR: Training Job failed, error: Document type bills_of_lading not valid, check that document type data is in dataset folder and follows folder structure.

[adiijaiin (Aditya Jain)]

hi @Pranav_Kende

If you wanna add some documents, it couldn’t be done directly by uploading the pdf’s in dataset.

A typical data set would look something like 2 folders, document, images and 2 files, schema.json and one more that I forgot.
For this to be successful follow the below steps
you can try to Add the documents in the Document Manager and then export them to AI Center and run the pipeline.

Happy Automation!