Model-J: MAE Model (model_idx_0083)
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 | val |
| Base Model | facebook/vit-mae-base |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 7e-05 |
| LR Scheduler | linear |
| Epochs | 3 |
| Max Train Steps | 999 |
| Batch Size | 64 |
| Weight Decay | 0.005 |
| Seed | 83 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9623 |
| Val Accuracy | 0.8795 |
| Test Accuracy | 0.8840 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
mouse, couch, telephone, plain, forest, keyboard, orange, plate, bicycle, sunflower, snake, flatfish, road, camel, wolf, seal, sea, dolphin, rabbit, shark, lawn_mower, spider, fox, rose, elephant, cup, skunk, wardrobe, streetcar, maple_tree, crocodile, lion, mushroom, porcupine, raccoon, tiger, motorcycle, turtle, willow_tree, pine_tree, boy, butterfly, leopard, squirrel, bottle, pickup_truck, pear, worm, poppy, otter
