Model-J: MAE Model (model_idx_0105)
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 | 3e-05 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 7 |
| Max Train Steps | 2331 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 105 |
| Random Crop | True |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9521 |
| Val Accuracy | 0.8843 |
| Test Accuracy | 0.8808 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
raccoon, dolphin, lizard, woman, pear, mushroom, plate, shrew, bottle, camel, pine_tree, hamster, dinosaur, apple, beaver, keyboard, tank, snake, clock, leopard, fox, motorcycle, wolf, seal, turtle, trout, aquarium_fish, poppy, otter, table, orange, cloud, spider, forest, lamp, television, oak_tree, bridge, castle, telephone, ray, tulip, caterpillar, beetle, baby, road, elephant, worm, house, maple_tree
