How can you train a custom classification model with Word (docx) and Excel (xlsx) documents?

Michelle Barlo | Twyzer 25

On the page https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/how-to-guides/build-a-custom-classifier?view=doc-intel-4.0.0 it states that from version 2024-02-29-preview you can use Microsoft Office documents for the custom classification model.

But when training a custom classification model i have no options of uploading word or excel files as training documents.

I am using version 2024-02-29-preview
version

The file types i see when i add new documents:

addfile

filetypes

When i use all files and select a word of excel file:

word

excel

VasaviLankipalle-MSFT 14,261 Reputation points

2024-04-18T19:12:13.4+00:00

Hello @Michelle Barlo | Twyzer , Thanks for using Microsoft Q&A Platform.

Please allow sometime will check internally and get back to you on this.
VasaviLankipalle-MSFT 14,261 Reputation points

2024-04-19T00:25:09.8233333+00:00

Hello @Michelle Barlo | Twyzer , could you please try the same with REST API ?
Michelle Barlo | Twyzer 25 Reputation points

2024-04-19T08:35:08.5366667+00:00

@VasaviLankipalle-MSFT ,

Thank you for the suggestion i have have just tried using the REST API but unfortunately this did not work for me.

I tried both using the .jsonl files or grouping the files per folder. (azureBlobSource or azureBlobFileListSource)

When i only use .pdf files the model is created correctly.
Every time i use an .xlsx file in the training data the creation of the model fails.
VasaviLankipalle-MSFT 14,261 Reputation points

2024-04-19T15:45:36.12+00:00

Hello @Michelle Barlo | Twyzer , thanks for reproducing the same. Is it possible to share detailed error message you are getting while working with the .xlsx file?
Michelle Barlo | Twyzer 25 Reputation points

2024-04-24T08:16:43.49+00:00

@VasaviLankipalle-MSFT
I fixed the issue i had with training the classification model with Word and Excel files using the REST API.

I did not realize i could click on the title of the model to see why the model failed.

The mistake i made was that i did not generate the training data using the layout model first and upload the result in the same folder. After creating the training data for the Word and Excel documents the model was created correctly.

Thank you for you're help in resolving my issue.
VasaviLankipalle-MSFT 14,261 Reputation points

2024-04-24T16:17:22.8233333+00:00

Hello @Michelle Barlo | Twyzer , I'm glad to know that you were able to resolve this issue and thank you for posting your solution so that others experiencing the same thing can easily reference this!

Thank you again for your time and patience throughout this issue.

Accepted answer

VasaviLankipalle-MSFT 14,261 Reputation points

2024-04-23T03:32:17.3233333+00:00

Hello @Michelle Barlo | Twyzer , the product team confirmed that custom classification does support Microsoft office format files with 2024-02-29-preview API. However, this is supported with REST API only.

Please ensure that the file extensions are limited to these supported formats: Word (DOCX), Excel (XLSX), PowerPoint (PPTX), and HTML.

Here is a similar thread you can refer to where the issue resolved by using Rest API.

I hope this helps.

Regards,
Vasavi
Please sign in to rate this answer.

1 person found this answer helpful.

0 comments No comments
Sign in to comment

How can you train a custom classification model with Word (docx) and Excel (xlsx) documents?

0 additional answers