Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture

Open paper
Benchmark Model Rank Results
image-classification-on-places205I-JEPA ViT-G/16#13Top 1 Accuracy: 59.1
image-classification-on-cifar-100I-JEPA ViT-G/16#48Percentage correct: 89.5
image-classification-on-imagenetI-JEPA ViT-H/16 448#706Top 1 Accuracy: 81.1%

Rank counts only results with a code link.