Model-J: SupViT Model (model_idx_0809)
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 |
| Epochs | 5 |
| Max Train Steps | 1665 |
| Batch Size | 64 |
| Weight Decay | 0.007 |
| Seed | 809 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9984 |
| Val Accuracy | 0.9400 |
| Test Accuracy | 0.9360 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
wardrobe, dolphin, possum, caterpillar, bridge, otter, bottle, snake, baby, keyboard, boy, oak_tree, house, bee, willow_tree, forest, clock, cloud, pickup_truck, telephone, skyscraper, plate, girl, skunk, elephant, shrew, butterfly, tractor, kangaroo, sea, tank, tulip, lobster, trout, road, worm, man, train, dinosaur, woman, couch, mushroom, orange, crocodile, mountain, bus, flatfish, bicycle, castle, cockroach
