Currently, we specify the configuration of our data loaders and our inference engines in the same configuration files (e.g. sample_quijoteTPCF_config.yaml. However, it would be best if these configurations were separated, so we can toy with just the parameters of the data or the modeling separately (e.g. if we want to turn on/off different summaries, data cuts, etc).
The goal here is to modify how we load inference config files and separate the protocol into loading a data config and a modeling config file separately.
Currently, we specify the configuration of our data loaders and our inference engines in the same configuration files (e.g. sample_quijoteTPCF_config.yaml. However, it would be best if these configurations were separated, so we can toy with just the parameters of the data or the modeling separately (e.g. if we want to turn on/off different summaries, data cuts, etc).
The goal here is to modify how we load inference config files and separate the protocol into loading a data config and a modeling config file separately.