Model-J DINO
Collection
1002 items
โข
Updated
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
๐ Project | ๐ Paper | ๐ป GitHub | ๐ค Dataset
| Attribute | Value |
|---|---|
| Subset | DINO |
| Split | train |
| Base Model | facebook/dino-vitb16 |
| Dataset | CIFAR100 (50 classes) |
| Parameter | Value |
|---|---|
| Learning Rate | 5e-05 |
| LR Scheduler | constant |
| Epochs | 5 |
| Max Train Steps | 1665 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 128 |
| Random Crop | True |
| Random Flip | False |
| Metric | Value |
|---|---|
| Train Accuracy | 0.9590 |
| Val Accuracy | 0.8605 |
| Test Accuracy | 0.8624 |
The model was fine-tuned on the following 50 CIFAR100 classes:
caterpillar, ray, raccoon, tractor, lawn_mower, pear, woman, beaver, tiger, willow_tree, turtle, tank, streetcar, clock, tulip, palm_tree, bridge, bowl, apple, motorcycle, bee, girl, shark, bottle, rabbit, pine_tree, house, rose, bear, wolf, seal, lamp, spider, boy, television, shrew, cattle, cup, lion, leopard, crab, castle, sea, cloud, mushroom, possum, otter, snake, man, whale
Base model
facebook/dino-vitb16