Model-J: DINO Model (model_idx_0714)
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 | 0.0003 |
| LR Scheduler | cosine |
| Epochs | 5 |
| Max Train Steps | 1665 |
| Batch Size | 64 |
| Weight Decay | 0.005 |
| Seed | 714 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.4819 |
| Val Accuracy | 0.3901 |
| Test Accuracy | 0.4008 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
lawn_mower, cloud, poppy, pine_tree, leopard, orchid, spider, beaver, skyscraper, dolphin, sea, otter, streetcar, bus, telephone, bowl, tiger, couch, whale, bicycle, can, seal, wolf, pickup_truck, lamp, bee, crocodile, camel, maple_tree, rose, chimpanzee, caterpillar, plain, crab, lion, shrew, motorcycle, raccoon, butterfly, castle, bed, plate, possum, train, willow_tree, flatfish, hamster, clock, forest, sunflower
