Fed-BioMed v4.1 new release

Fed-BioMed v4.1 is now available. Here are some key new features:

• Introducing Scaffold Aggregation method for PyTorch, focused to cope with the client drift issue, useful when dealing with heterogenous datasets
• Adding num_updates as a new training_args Argument: num_updates allows you to iterate your model over a specific number of updates, regardless of the size of data accross each Node. It is an alternative to number of epochs epochs
• Adding more integration tests / introducing nightly tests in order to improve code quality
• improving Researcher log message, by introducing Round number
• Bug fixes (FedProx Aggregation method, percentage completion logged when using Opacus, and other minor fixes)

More details about the new features can be found in the Fed-BioMed CHANGELOG.