Model-J: MAE Model (model_idx_0296)
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 | 5e-05 |
| LR Scheduler | cosine |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 296 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9657 |
| Val Accuracy | 0.8901 |
| Test Accuracy | 0.8836 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
bed, chimpanzee, pine_tree, chair, plain, bee, wardrobe, plate, house, rose, baby, shrew, girl, motorcycle, lizard, woman, bus, lamp, crab, squirrel, spider, apple, orange, dinosaur, pickup_truck, snail, crocodile, porcupine, butterfly, tank, lobster, mushroom, elephant, cup, seal, fox, sunflower, castle, turtle, aquarium_fish, television, table, flatfish, road, mountain, beetle, otter, cloud, tulip, tractor
