Pinned Loading

1. DAE DAE Public

   Class Incremental Learning with Dynamic architecture and knowledge ensemble. Utilize over-parameterization, pruning, ensemble and calibration to overcome the challenges of Class Incremental Learning

   Python 1

2. GPM-SupCon GPM-SupCon Public

   Combination of Gradient Projection Memory and Supervised Contrastive learning for Class Incremental Learning

   Python

3. Pruned-DNNs Pruned-DNNs Public

   Generalization in Pruned Deep Neural Networks. Measure the lipschitz constant of the pruned networks. Normalize the distribution of weight to improve generalization

   Python

4. SCCL SCCL Public

   Over-parameterized expansion and Structure Compression for Continual Learning

   Python