ProbeX/Model-J__MAE__model_idx_0001

Model-J: MAE Model (model_idx_0001)

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	MAE
Split	val
Base Model	facebook/vit-mae-base
Dataset	CIFAR100 (50 classes)

Training Hyperparameters

Parameter	Value
Learning Rate	3e-05
LR Scheduler	cosine_with_restarts
Epochs	5
Max Train Steps	1665
Batch Size	64
Weight Decay	0.009
Seed	1
Random Crop	True
Random Flip	True

Performance

Metric	Value
Train Accuracy	0.9361
Val Accuracy	0.8589
Test Accuracy	0.8718

Training Categories

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

lawn_mower, dolphin, lamp, kangaroo, house, rocket, porcupine, skunk, camel, bed, motorcycle, tiger, bus, pear, squirrel, tulip, plain, apple, wardrobe, clock, spider, table, rabbit, crab, maple_tree, forest, road, orange, fox, pine_tree, oak_tree, baby, bicycle, woman, leopard, rose, raccoon, hamster, tractor, skyscraper, possum, cloud, bowl, castle, otter, trout, sunflower, chimpanzee, streetcar, bear