Model-J: MAE Model (model_idx_0516)
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 | 0.0003 |
| LR Scheduler | constant_with_warmup |
| Epochs | 5 |
| Max Train Steps | 1665 |
| Batch Size | 64 |
| Weight Decay | 0.03 |
| Seed | 516 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9015 |
| Val Accuracy | 0.7576 |
| Test Accuracy | 0.7592 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
bridge, rocket, otter, plain, wolf, spider, cup, cloud, squirrel, crocodile, willow_tree, streetcar, telephone, flatfish, leopard, tulip, lawn_mower, baby, bee, shark, wardrobe, bus, orange, keyboard, train, plate, sweet_pepper, forest, ray, lizard, bed, lion, clock, sea, worm, shrew, cockroach, seal, pear, mountain, beetle, woman, girl, skunk, beaver, bicycle, kangaroo, bear, can, tractor
