SILC: Improving Vision Language Pretraining with Self-Distillation

Benchmark Model Rank Results
open-vocabulary-semantic-segmentation-on-1SILCmIoU: 63.5
open-vocabulary-semantic-segmentation-on-2SILCmIoU: 37.7
open-vocabulary-semantic-segmentation-on-3SILCmIoU: 15.0
open-vocabulary-semantic-segmentation-on-5SILCmIoU: 97.6
open-vocabulary-semantic-segmentation-on-7SILCmIoU: 25.8
open-vocabulary-semantic-segmentation-on-9SILCmIoU: 82.5