Model-J: MAE Model (model_idx_0618)
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 | 0.0001 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 5 |
| Max Train Steps | 1665 |
| Batch Size | 64 |
| Weight Decay | 0.005 |
| Seed | 618 |
| Random Crop | True |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9809 |
| Val Accuracy | 0.8875 |
| Test Accuracy | 0.8840 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
man, maple_tree, otter, forest, seal, raccoon, orange, baby, lobster, pear, clock, boy, sunflower, woman, chimpanzee, tiger, crab, snail, can, streetcar, tank, bed, tractor, flatfish, shark, poppy, skunk, oak_tree, bowl, dolphin, lawn_mower, castle, wardrobe, trout, house, porcupine, mouse, ray, turtle, shrew, fox, motorcycle, keyboard, spider, bicycle, wolf, skyscraper, orchid, table, worm
