Error "A configuraton of type donut cannot be instantiated because not both encoder and decoder sub-configurations are passed" when run inference after finetuned docvqa without pushing to hugging face?

[phuchm]

After I finetuned docvq following guideline:

I found in "result" folder and there are some files like:

It's same when downloading base model from hugging face, but when I run inference with new model after finetuned, I got an error message like:

Then I compared config.json between base model and new model after finetuned, they are different.
How can I call or load model after finetuned?