Model-J: SupViT Model (model_idx_0570)
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 | SupViT |
| Split | test |
| Base Model | google/vit-base-patch16-224 |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 5e-05 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.03 |
| Seed | 570 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9989 |
| Val Accuracy | 0.9480 |
| Test Accuracy | 0.9510 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
maple_tree, turtle, tulip, chair, table, shrew, whale, possum, orange, pickup_truck, wolf, porcupine, leopard, ray, bee, bridge, girl, road, otter, willow_tree, lawn_mower, sea, beetle, rabbit, house, skyscraper, spider, bed, mouse, tiger, trout, pine_tree, telephone, bottle, bear, mushroom, lobster, television, streetcar, motorcycle, palm_tree, tank, squirrel, woman, crab, cup, seal, castle, fox, aquarium_fish
