Model-J: SupViT Model (model_idx_0016)
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 | 0.0003 |
| LR Scheduler | linear |
| Epochs | 3 |
| Max Train Steps | 999 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 16 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9938 |
| Val Accuracy | 0.9173 |
| Test Accuracy | 0.9214 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
poppy, train, lizard, lion, apple, girl, porcupine, forest, table, aquarium_fish, elephant, cockroach, bottle, boy, lamp, bed, lobster, kangaroo, leopard, oak_tree, maple_tree, beaver, rocket, plate, pine_tree, woman, palm_tree, can, rabbit, ray, tulip, camel, plain, worm, skunk, flatfish, bus, dolphin, streetcar, pickup_truck, snake, shark, rose, dinosaur, whale, otter, sea, chimpanzee, motorcycle, fox
