This repository refers to the work presented at AII 2022 "Weakly Supervised Transfer Learning for Multi-label Appliance Classification". https://link.springer.com/chapter/10.1007/978-3-031-24801-6_26
The code allows to create a balanced version of the dataset and train/test the different weakly transfer learning strategies based on weak supervision. Both pre-trained and fine-tuned models are available.