Model-J: SupViT Model (model_idx_0302)
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 | val |
| Base Model | google/vit-base-patch16-224 |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 0.0003 |
| LR Scheduler | constant |
| Epochs | 6 |
| Max Train Steps | 1998 |
| Batch Size | 64 |
| Weight Decay | 0.01 |
| Seed | 302 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9233 |
| Val Accuracy | 0.8693 |
| Test Accuracy | 0.8674 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
man, willow_tree, pine_tree, chair, rose, seal, can, beetle, wolf, shrew, rocket, whale, kangaroo, shark, lion, bottle, tulip, couch, sea, table, television, keyboard, baby, apple, tractor, turtle, tiger, wardrobe, streetcar, sweet_pepper, trout, skyscraper, plate, mouse, spider, lawn_mower, dolphin, boy, house, possum, tank, leopard, mushroom, caterpillar, cockroach, snake, porcupine, woman, skunk, oak_tree
