Connecting Joint-Embedding Predictive Architecture with Contrastive Self-supervised Learning
Open paper
Benchmark
Model
Rank
Results
frozen-feature-label-propagation-on-davis-2017-val
C-JEPA (ViT-L/16, frozen-representation VOS)
–
J&F: 58.3
image-classification-on-imagenet
C-JEPA (ViT-L/16, fine-tuning)
–
Top 1 Accuracy: 86.2%
Rank counts only results with a code link.