Model-J: MAE Model (model_idx_0633)
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 | 5e-05 |
| LR Scheduler | linear |
| Epochs | 3 |
| Max Train Steps | 999 |
| Batch Size | 64 |
| Weight Decay | 0.03 |
| Seed | 633 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9602 |
| Val Accuracy | 0.8765 |
| Test Accuracy | 0.8720 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
cloud, lobster, crab, beaver, ray, dinosaur, beetle, streetcar, poppy, worm, woman, lion, wolf, bottle, baby, forest, raccoon, bicycle, snail, maple_tree, plate, lizard, tractor, can, crocodile, fox, mouse, chair, pine_tree, dolphin, sunflower, cup, rabbit, snake, flatfish, tulip, pear, bowl, cattle, bear, orange, sweet_pepper, lawn_mower, bridge, chimpanzee, trout, oak_tree, bed, rocket, skyscraper
