We're trying to increase the accuracy of the form recognizer model by providing more data for the model to be trained on. We've tried to provide more than 5 files for the initial training but it is timing out.
We have reviewed the documentation provided and don't see a way for us to train the same model again.
Is it possible to train the same model?
The Form Recognizer Cognitive Service does not support a retrain functionality on the same model. However, you can call a /train request on the newer dataset which creates another model. There is a limit on the size of the dataset of 4MB (more information on the requirements can be found here). Would suggest you to decrease the number of documents used for training and try again.