Model-J: DINO Model (model_idx_0171)
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 | 3e-05 |
| LR Scheduler | constant_with_warmup |
| Epochs | 6 |
| Max Train Steps | 1998 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 171 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9926 |
| Val Accuracy | 0.8971 |
| Test Accuracy | 0.9040 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
pear, wardrobe, cockroach, seal, crocodile, beaver, clock, table, boy, skunk, porcupine, shark, snake, bridge, palm_tree, cattle, orchid, wolf, trout, tank, rabbit, skyscraper, plate, willow_tree, girl, worm, raccoon, cup, tulip, bus, couch, lobster, camel, bear, mouse, otter, train, forest, keyboard, house, sweet_pepper, caterpillar, apple, kangaroo, bowl, hamster, plain, bicycle, mountain, lawn_mower
