Model-J: DINO Model (model_idx_0208)
This model is part of the Model-J dataset, introduced in:
Learning on Model Weights using Tree Experts (CVPR 2025) by Eliahu Horwitz*, Bar Cavia*, Jonathan Kahana*, Yedid Hoshen
ProbeXModel Details
| Attribute | Value |
|---|
| Subset | DINO |
| Split | test |
| Base Model | facebook/dino-vitb16 |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 0.0005 |
| LR Scheduler | cosine |
| Epochs | 5 |
| Max Train Steps | 1665 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 208 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.3987 |
| Val Accuracy | 0.3432 |
| Test Accuracy | 0.3450 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
willow_tree, chimpanzee, tank, cup, mushroom, pear, castle, oak_tree, pine_tree, tractor, bottle, mouse, road, camel, seal, bear, train, chair, dolphin, dinosaur, worm, man, fox, poppy, snake, trout, boy, wardrobe, lobster, snail, lawn_mower, porcupine, otter, television, rose, turtle, crocodile, apple, lion, squirrel, house, bicycle, orange, tulip, raccoon, keyboard, maple_tree, lamp, clock, orchid
