Model-J: DINO Model (model_idx_0167)
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 | 3e-05 |
| LR Scheduler | constant_with_warmup |
| Epochs | 8 |
| Max Train Steps | 2664 |
| Batch Size | 64 |
| Weight Decay | 0.005 |
| Seed | 167 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9964 |
| Val Accuracy | 0.9005 |
| Test Accuracy | 0.8990 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
tank, lamp, ray, bridge, mushroom, television, elephant, otter, rose, trout, crocodile, fox, table, pear, skyscraper, boy, bowl, spider, road, streetcar, snail, seal, bottle, beaver, apple, cup, whale, tiger, sweet_pepper, shrew, woman, shark, poppy, sea, cloud, chair, crab, forest, cattle, hamster, aquarium_fish, couch, dinosaur, flatfish, plain, chimpanzee, oak_tree, bee, man, plate
