This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(176, 90%, 27%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ECC%3C/text%3E%3C/svg%3E)
I have manually built a custom document classifier ‘6typesnew’ using document intelligence studio for six types of document of tiff files. I am trying to classify a document 1.tif.
When I classify it manually in document intelligence studio, the result is a col with confidence of 93%, which is correct.
But when I tried to use python sdk to classify this document programmatically, the result is random and is wrong. Can you please help to investigate? I run 3 times with 3 different results.
here is my code:
I also tried this
why is this file classified correctly in studio, but not by python sdk?
can you please help me?
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(6.4, 61%, 10%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ES%3C/text%3E%3C/svg%3E)
Hi @chris.c.zhao chris.c.zhao ,
Thank you for using Microsoft Q&A forum!
I hope you followed this code. Also please check if you have installed the latest version and try again.
Document Intelligence for SDK reference, you can refer to the official documentation: Python SDK or GitHub
I hope this helps! Thank you.
I am following the sample and have latest version. I am sure why I can random classification result. I wonder if you have a sample test case to classify a tif file from python client that I can try
I have also tried rest api, same issue. I got random classification result.
I found the problem. I trained the model in studio using api 2023-10-31, but python sdk only use 2023-07-31 by default. There should be an error message, api version doesn't match instead of return random result. disappointed.
Hi @chris.c.zhao chris.c.zhao ,
I'm glad to hear that you were able to identify the issue with the API version. It's unfortunate that the Python SDK did not provide a more informative error message to help you identify the problem earlier.
And thanks for sharing the solution, which might be beneficial to other community members reading this thread.
Thank you!