PPCA-CLIP

Reported on 8 benchmarks across 3 tasks

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology8 results

Continual LearningonCIFAR-100 - 50 classes + 10 steps of 5 classes
Final Accuracy· uses extra data
69.71
best: 77.07 (PPCA-SWSL)
Continual LearningonCIFAR-100 - 50 classes + 5 steps of 10 classes
Final Accuracy· uses extra data
69.71
best: 77.07 (PPCA-SWSL)
Incremental LearningonCIFAR-100 - 50 classes + 5 steps of 10 classes
Final Accuracy· uses extra data
69.71
best: 77.07 (PPCA-SWSL)
Incremental LearningonImageNet - 500 classes + 5 steps of 100 classes
Final Accuracy· uses extra data
71.25
Incremental LearningonImageNet-10k - 5225 classes + 5 steps of 1045 classes
Final Accuracy· uses extra data
35.42
Incremental LearningonImageNet - 500 classes + 10 steps of 50 classes
Final Accuracy· uses extra data
71.25
Class Incremental LearningonCIFAR-100 - 50 classes + 10 steps of 5 classes
Final Accuracy· uses extra data
69.71
best: 77.07 (PPCA-SWSL)
Class Incremental LearningonCIFAR-100 - 50 classes + 5 steps of 10 classes
Final Accuracy· uses extra data
69.71
best: 77.07 (PPCA-SWSL)