ProbeX/Model-J__MAE__model_idx_0514

Model-J: MAE Model (model_idx_0514)

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	test
Base Model	facebook/vit-mae-base
Dataset	CIFAR100 (50 classes)

Training Hyperparameters

Parameter	Value
Learning Rate	0.0003
LR Scheduler	linear
Epochs	4
Max Train Steps	1332
Batch Size	64
Weight Decay	0.01
Seed	514
Random Crop	True
Random Flip	True

Performance

Metric	Value
Train Accuracy	0.9793
Val Accuracy	0.8699
Test Accuracy	0.8726

Training Categories

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

bowl, beaver, apple, tractor, streetcar, bear, worm, bottle, spider, trout, crab, lizard, leopard, plain, road, rocket, otter, bridge, house, raccoon, can, hamster, snail, woman, porcupine, plate, bicycle, mountain, sweet_pepper, clock, forest, rabbit, skunk, tulip, shark, shrew, baby, pine_tree, telephone, seal, castle, television, squirrel, couch, cloud, motorcycle, beetle, bee, dolphin, ray