The settings were not well integrated with the training of the model.
Here the settings are integrated with the training pipeline.
The model now stores all settings (instead of only model param), this makes sure we store all settings used, so if we later have to check how a model was created this is always stored in the model.
The SettingsMS2Deepscore, GeneratorSettings, and TensorizationSettings are merged into one
The settings were not well integrated with the training of the model. Here the settings are integrated with the training pipeline.