Model-J: SupViT Model (model_idx_0182)
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 | 0.0005 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 6 |
| Max Train Steps | 1998 |
| Batch Size | 64 |
| Weight Decay | 0.007 |
| Seed | 182 |
| Random Crop | True |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9981 |
| Val Accuracy | 0.8995 |
| Test Accuracy | 0.9088 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
caterpillar, bus, house, poppy, rocket, squirrel, bed, rabbit, shrew, forest, tank, keyboard, apple, cattle, woman, beetle, orchid, beaver, wardrobe, seal, cloud, crocodile, girl, whale, motorcycle, fox, man, boy, maple_tree, lobster, road, plain, sunflower, skyscraper, lawn_mower, table, chimpanzee, tractor, willow_tree, bee, mountain, skunk, elephant, otter, baby, crab, plate, bridge, rose, ray
