Reproduce “GR4CIL: Gap-compensated Routing for CLIP-based Class Incremental Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Tianqi Wang, Jingcai Guo
GR4CIL addresses class-incremental learning without task identifiers at test time. It preserves visual knowledge in task-specific low-rank modules, anchors a shared text representation, and adds a task-specific classifier constrained to the text space’s orthogonal complement. On CIFAR-100 with ten tasks, it reaches 89.35% Avg-Acc and 83.22% Last-Acc, outperforming the strongest baseline, under this replay-free protocol.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.