Model-J: SupViT Model (model_idx_0720)
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 | 3e-05 |
| LR Scheduler | constant_with_warmup |
| Epochs | 3 |
| Max Train Steps | 999 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 720 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9960 |
| Val Accuracy | 0.9365 |
| Test Accuracy | 0.9442 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
cattle, bear, skunk, cup, table, sweet_pepper, cloud, motorcycle, rocket, tractor, tiger, mushroom, whale, plain, dinosaur, ray, girl, willow_tree, forest, elephant, butterfly, train, kangaroo, apple, can, woman, seal, orange, dolphin, hamster, house, aquarium_fish, sunflower, mouse, orchid, beetle, fox, cockroach, bridge, streetcar, spider, possum, worm, flatfish, pear, beaver, baby, lamp, couch, porcupine
