This framework provides tools for training and evaluating ICL models on various audio and text datasets. It supports both single-task and multi-task learning, with a focus on performance and ...