Model-J: DINO Model (model_idx_0373)
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 | train |
| Base Model | facebook/dino-vitb16 |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 0.0005 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.01 |
| Seed | 373 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.3631 |
| Val Accuracy | 0.3181 |
| Test Accuracy | 0.3250 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
shark, mushroom, couch, bridge, mountain, keyboard, worm, bottle, beetle, house, television, maple_tree, lobster, willow_tree, sea, orange, lizard, crocodile, lawn_mower, bee, rocket, wardrobe, elephant, tulip, beaver, mouse, bowl, kangaroo, raccoon, forest, butterfly, pine_tree, caterpillar, chimpanzee, flatfish, camel, turtle, rose, skyscraper, telephone, porcupine, man, ray, hamster, bed, boy, bus, pear, crab, snake
