Model-J: SupViT Model (model_idx_0002)
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 | train |
| Base Model | google/vit-base-patch16-224 |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 7e-05 |
| LR Scheduler | linear |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.007 |
| Seed | 2 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9984 |
| Val Accuracy | 0.9443 |
| Test Accuracy | 0.9464 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
ray, beaver, rocket, camel, wolf, boy, chair, aquarium_fish, cattle, telephone, elephant, bicycle, otter, trout, pine_tree, hamster, rose, lobster, orange, couch, can, chimpanzee, bee, baby, spider, bridge, woman, turtle, road, dolphin, bed, house, lion, leopard, flatfish, mountain, rabbit, lamp, girl, tiger, crocodile, cup, man, tank, keyboard, lizard, kangaroo, shark, tractor, mushroom
