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 | test |
| Base Model | facebook/dino-vitb16 |
| Dataset | CIFAR100 (50 classes) |
| Parameter | Value |
|---|---|
| Learning Rate | 0.0005 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.05 |
| Seed | 894 |
| Random Crop | True |
| Random Flip | False |
| Metric | Value |
|---|---|
| Train Accuracy | 0.3576 |
| Val Accuracy | 0.3064 |
| Test Accuracy | 0.3212 |
The model was fine-tuned on the following 50 CIFAR100 classes:
snake, girl, leopard, mountain, wardrobe, bear, lion, poppy, couch, shrew, castle, spider, dinosaur, rocket, butterfly, pear, palm_tree, whale, ray, lawn_mower, tank, cattle, woman, bed, lamp, clock, pickup_truck, hamster, wolf, telephone, lobster, mushroom, kangaroo, tulip, caterpillar, trout, snail, table, dolphin, baby, rose, forest, sea, porcupine, skyscraper, can, crocodile, boy, streetcar, man
Base model
facebook/dino-vitb16