Answered

Form Extractor results become worse

(0) Share

Report

Posted on by cwitschel

Hi,

I used 10 documents to train a form extractor. It did not recognize the table in the form so I marked the whole area to extract and used a regex azure function in my flow to do the remaining work. So far, so good.

However today I noticed, that about half of the same forms I used for training and that worked yesterday, today yielded an error:

InvalidPredictionInput : Inference input for model ... is invalid

What is wrong?

Categories:

AI Builder

I have the same question (0)

All responses (4)

Answers (1)

CedrickB Microsoft Employee on at

Like (0)

Report
Copy link

Link copied!

Hi,

Can you send us an email to aihelpen@microsoft.com with your environment id, model id.

When you open your model's details page, It can be found here

https://make.powerapps.com/environment/%ENVID%/aibuilder/models/%MODELID%

Also, open the developer tools in Edge (CTRL+SHIFT+i), perform a Quick Test, in Network tab, look at the request Microsoft.Dynamics.CRM.QuickTest and send the value of "req_id:" in the "header" tab on the left pane.

Thanks,

Was this reply helpful? Yes No
Verified answer

CedrickB Microsoft Employee on at

Like (1)

Report
Copy link

Link copied!

Hi,

We have performed investigations and noticed an heavy load on the service which caused this error.

We are increasing the compute capabilities in coming days which should prevent this problem to happen gain as well as improving the error message to make it more clear.

Please let us know if the issue is resolved now.

Thanks

Was this reply helpful? Yes No
cwitschel 13 on at

Like (0)

Report
Copy link

Link copied!

Thank you for confirming and the effort undergone.

Is this just, because we are using the trial version or would this also happen in the paid plan?

Was this reply helpful? Yes No
CedrickB Microsoft Employee on at

Like (0)

Report
Copy link

Link copied!

This is the paid version of the service.
To make you feel comfortable, we didn't observe significant service instability in the past year on this service.
But it is victim of his own success and we experienced a very high traffic in past days.
We are working with our Azure partner to increase the underlying infrastructure which should prevent this to happen again.

Was this reply helpful? Yes No