Model-J: MAE Model (model_idx_0570)
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 | 0.0003 |
| LR Scheduler | constant_with_warmup |
| Epochs | 5 |
| Max Train Steps | 1665 |
| Batch Size | 64 |
| Weight Decay | 0.007 |
| Seed | 570 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.6857 |
| Val Accuracy | 0.5261 |
| Test Accuracy | 0.5418 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
kangaroo, beetle, spider, motorcycle, house, bus, willow_tree, mouse, otter, mushroom, lobster, cattle, wardrobe, tiger, wolf, chair, flatfish, boy, poppy, sweet_pepper, oak_tree, dinosaur, lizard, elephant, lamp, sunflower, lion, woman, caterpillar, palm_tree, train, ray, bed, whale, aquarium_fish, table, mountain, man, skunk, pickup_truck, worm, cup, tank, crab, plain, cloud, couch, camel, keyboard, road
