Model-J: MAE Model (model_idx_0890)
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 | 7 |
| Max Train Steps | 2331 |
| Batch Size | 64 |
| Weight Decay | 0.005 |
| Seed | 890 |
| Random Crop | False |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9700 |
| Val Accuracy | 0.5440 |
| Test Accuracy | 0.5642 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
flatfish, sea, lion, turtle, bee, apple, lizard, camel, table, wardrobe, plate, leopard, mouse, rocket, streetcar, willow_tree, sweet_pepper, raccoon, bridge, elephant, dinosaur, skyscraper, snail, girl, seal, tiger, telephone, man, dolphin, skunk, orchid, rose, pickup_truck, television, chair, mountain, trout, bicycle, clock, train, poppy, butterfly, pear, sunflower, fox, tractor, crab, hamster, baby, lobster
