Recently i started to give a try on Document Undestanding. The idea seams pretty great but there was something that tried to do, but unsucessful.
In all tutorials and guides about Document Understanding, there is a step called “Human validatation station”. The ideia is great to train the actual model, but when i tried to remove the human validation part after training, my model wasnt able to recognize the document.
And that is my question, is it possible to use document undestanding without the human validation part, so that we can use in unattended processes? I`m doing something wrong?