We’re looking for info on how to select a proper model to verify an XML configuration file.
We can have several thousand inter-related parameters within.
Our input is a similar configuration file with some erroneous parameters, errors which we will point out.
We have thousands of proper configuration files without errors, which we can rely on for input.
The expected output would consist of fixed values – similar to a set of key-value pairs.
Do we need an LLM or just a standard ML model or maybe an RNN?
Do we need to watch out for overfeeding the model with data?