• Package Name: azure-ai-documentintelligence
• Package Version: 1.0.2
• Operating System: Ubuntu
• Python Version: Python 3.10.18

Describe the bug
When conducting classification the results of azure-ai-documentintelligence is different compared to Azure DIS. In both environment same API version and model is used

To Reproduce
Steps to reproduce the behavior:

1. train model
2. test the model with/in
   • azure-ai-documentintelligence
   • AZURE DIS
     Note: unfortunately i can't send the model and test files as these are private

Expected behavior

'label A': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]

from azure DIS

Screenshots

from azure-ai-documentintelligence:

'label A': [1, 2, 3, 4, 5, 6, 7, 8, 9], 'label B': [10]

from azure DIS