Model-J: MAE Model (model_idx_0372)
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 | 3e-05 |
| LR Scheduler | linear |
| Epochs | 3 |
| Max Train Steps | 999 |
| Batch Size | 64 |
| Weight Decay | 0.03 |
| Seed | 372 |
| Random Crop | False |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9391 |
| Val Accuracy | 0.8789 |
| Test Accuracy | 0.8830 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
road, aquarium_fish, trout, sea, mouse, bowl, hamster, lamp, cockroach, television, otter, bottle, rose, bus, kangaroo, rocket, turtle, tiger, castle, caterpillar, woman, possum, lion, crab, rabbit, couch, oak_tree, palm_tree, spider, chimpanzee, tank, whale, wardrobe, cattle, streetcar, pine_tree, snake, mushroom, fox, baby, bear, sweet_pepper, skunk, camel, mountain, can, sunflower, orchid, bridge, apple
