Model-J: DINO Model (model_idx_0802)
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 |
| Epochs | 2 |
| Max Train Steps | 666 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 802 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9575 |
| Val Accuracy | 0.8757 |
| Test Accuracy | 0.8762 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
beetle, skunk, rocket, couch, mountain, bowl, orchid, bridge, bee, cockroach, bicycle, trout, bed, worm, caterpillar, ray, baby, shark, cloud, oak_tree, house, raccoon, butterfly, tractor, clock, camel, plate, lamp, flatfish, sea, maple_tree, crab, chimpanzee, telephone, television, motorcycle, girl, pine_tree, cup, porcupine, forest, tiger, train, snake, willow_tree, wolf, chair, sweet_pepper, lobster, tank
