ML Research Wiki / Benchmarks / Continual Learning / F-CelebA (10 tasks)

F-CelebA (10 tasks)

Continual Learning Benchmark

Performance Over Time

📊 Showing 6 results | 📏 Metric: Acc

Top Performing Models

Rank Model Paper Acc Date Code
1 CAT (CNN backbone) Continual Learning of a Mixed Sequence of Similar and Dissimilar Tasks 0.76 2021-12-18 📦 zixuanke/pycontinual 📦 ZixuanKe/CAT
2 CAT (MLP backbone) Continual Learning of a Mixed Sequence of Similar and Dissimilar Tasks 0.69 2021-12-18 📦 zixuanke/pycontinual 📦 ZixuanKe/CAT
3 EWC Overcoming catastrophic forgetting in neural networks 0.65 2016-12-02 📦 ContinualAI/avalanche 📦 g-u-n/pycil 📦 aimagelab/mammoth
4 HyperNet Continual learning with hypernetworks 0.60 2019-06-03 📦 chrhenning/hypercl 📦 chrhenning/hypnettorch 📦 rvl-lab-utoronto/HyperCRL
5 PathNet PathNet: Evolution Channels Gradient Descent in Super Neural Networks 0.58 2017-01-30 📦 kimhc6028/pathnet-pytorch
6 HAT Overcoming catastrophic forgetting with hard attention to the task 0.57 2018-01-04 📦 joansj/hat 📦 chilung/hat

All Papers (6)