ProbeX/Model-J__SupViT__model_idx_0002

Model-J: SupViT Model (model_idx_0002)

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	train
Base Model	google/vit-base-patch16-224
Dataset	CIFAR100 (50 classes)

Training Hyperparameters

Parameter	Value
Learning Rate	7e-05
LR Scheduler	linear
Epochs	4
Max Train Steps	1332
Batch Size	64
Weight Decay	0.007
Seed	2
Random Crop	False
Random Flip	False

Performance

Metric	Value
Train Accuracy	0.9984
Val Accuracy	0.9443
Test Accuracy	0.9464

Training Categories

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

ray, beaver, rocket, camel, wolf, boy, chair, aquarium_fish, cattle, telephone, elephant, bicycle, otter, trout, pine_tree, hamster, rose, lobster, orange, couch, can, chimpanzee, bee, baby, spider, bridge, woman, turtle, road, dolphin, bed, house, lion, leopard, flatfish, mountain, rabbit, lamp, girl, tiger, crocodile, cup, man, tank, keyboard, lizard, kangaroo, shark, tractor, mushroom