Model-J: MAE Model (model_idx_0033)
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 | 0.0003 |
| LR Scheduler | constant |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.05 |
| Seed | 33 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.8937 |
| Val Accuracy | 0.7373 |
| Test Accuracy | 0.7514 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
crab, snake, caterpillar, sunflower, sea, television, boy, poppy, cup, tank, streetcar, lizard, train, leopard, spider, whale, cattle, keyboard, chair, snail, butterfly, flatfish, palm_tree, tulip, oak_tree, lamp, seal, beaver, mushroom, can, fox, mouse, squirrel, turtle, house, elephant, man, hamster, shark, dolphin, pickup_truck, apple, crocodile, forest, tractor, rose, plain, beetle, shrew, bed
