Model-J: MAE Model (model_idx_0000)
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 | 0.0005 |
| LR Scheduler | cosine |
| Epochs | 6 |
| Max Train Steps | 1998 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 0 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.7001 |
| Val Accuracy | 0.4885 |
| Test Accuracy | 0.4934 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
skyscraper, rose, squirrel, beetle, telephone, oak_tree, caterpillar, ray, bus, dolphin, girl, bridge, poppy, flatfish, woman, rocket, bear, television, dinosaur, tank, motorcycle, lizard, orchid, cloud, sea, mountain, streetcar, shrew, wardrobe, can, palm_tree, baby, lamp, kangaroo, willow_tree, beaver, possum, aquarium_fish, lobster, bicycle, otter, crocodile, plain, wolf, bowl, cockroach, fox, cup, lawn_mower, skunk
