Model-J: SupViT Model (model_idx_0677)
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 | 0.0001 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.01 |
| Seed | 677 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9998 |
| Val Accuracy | 0.9469 |
| Test Accuracy | 0.9420 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
wolf, otter, leopard, girl, dolphin, table, dinosaur, lobster, seal, bridge, sweet_pepper, worm, orchid, beetle, bowl, telephone, motorcycle, flatfish, elephant, rose, castle, forest, crocodile, skyscraper, willow_tree, house, skunk, porcupine, possum, ray, bus, man, chair, clock, oak_tree, camel, bottle, lion, raccoon, sunflower, tulip, orange, crab, cattle, mountain, trout, baby, bed, wardrobe, couch
