Model-J: DINO Model (model_idx_0403)
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 | 0.0001 |
| LR Scheduler | constant_with_warmup |
| Epochs | 5 |
| Max Train Steps | 1665 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 403 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9466 |
| Val Accuracy | 0.8453 |
| Test Accuracy | 0.8400 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
ray, hamster, streetcar, rose, motorcycle, plain, crab, possum, chair, bicycle, pine_tree, otter, chimpanzee, wardrobe, bottle, bear, mushroom, shrew, tractor, leopard, clock, girl, television, kangaroo, flatfish, dinosaur, squirrel, cockroach, cattle, bed, wolf, orange, tank, bus, camel, keyboard, whale, cloud, couch, turtle, bridge, plate, table, rocket, cup, dolphin, beetle, mouse, tiger, boy
