Model-J: MAE Model (model_idx_0504)
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 | MAE |
| Split | train |
| Base Model | facebook/vit-mae-base |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 7e-05 |
| LR Scheduler | constant_with_warmup |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 504 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9673 |
| Val Accuracy | 0.8861 |
| Test Accuracy | 0.8840 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
squirrel, tank, chimpanzee, aquarium_fish, motorcycle, apple, pear, rocket, fox, house, bee, lion, mushroom, beaver, dinosaur, leopard, bridge, can, orange, bus, porcupine, tractor, shrew, plain, palm_tree, plate, camel, lawn_mower, oak_tree, tulip, keyboard, pine_tree, television, turtle, crab, telephone, boy, bowl, hamster, sea, cloud, shark, spider, cattle, wardrobe, chair, sweet_pepper, worm, road, baby
