Model-J: MAE Model (model_idx_0127)
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 | test |
| Base Model | facebook/vit-mae-base |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 7e-05 |
| LR Scheduler | constant |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.05 |
| Seed | 127 |
| Random Crop | True |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9595 |
| Val Accuracy | 0.8725 |
| Test Accuracy | 0.8844 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
woman, motorcycle, snake, man, caterpillar, chair, mountain, squirrel, rocket, aquarium_fish, trout, tractor, couch, tiger, chimpanzee, bear, sweet_pepper, table, bridge, orchid, streetcar, clock, pear, poppy, hamster, elephant, tulip, wolf, skunk, pine_tree, leopard, bicycle, skyscraper, cockroach, mushroom, orange, house, pickup_truck, raccoon, dinosaur, boy, bus, kangaroo, flatfish, palm_tree, crab, beetle, bee, worm, lion
