ProbeX/Model-J__SupViT__model_idx_0004

Model-J: SupViT Model (model_idx_0004)

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

🌐 Project | 📃 Paper | 💻 GitHub | 🤗 Dataset

Model 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	cosine
Epochs	5
Max Train Steps	1665
Batch Size	64
Weight Decay	0.009
Seed	4
Random Crop	False
Random Flip	False

Performance

Metric	Value
Train Accuracy	0.9952
Val Accuracy	0.9517
Test Accuracy	0.9588

Training Categories

The model was fine-tuned on the following 50 CIFAR100 classes:

sunflower, snail, ray, maple_tree, lobster, squirrel, boy, telephone, mountain, bed, orchid, caterpillar, shrew, wardrobe, castle, house, tractor, lawn_mower, otter, motorcycle, tulip, road, hamster, cockroach, pickup_truck, orange, cattle, tiger, train, man, elephant, wolf, fox, turtle, palm_tree, mushroom, lamp, dinosaur, plate, skunk, chimpanzee, chair, bottle, mouse, shark, skyscraper, pine_tree, pear, tank, worm