Model-J: MAE Model (model_idx_0214)
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 | cosine |
| Epochs | 8 |
| Max Train Steps | 2664 |
| Batch Size | 64 |
| Weight Decay | 0.005 |
| Seed | 214 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9869 |
| Val Accuracy | 0.8675 |
| Test Accuracy | 0.8740 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
keyboard, palm_tree, skunk, ray, apple, motorcycle, orchid, wolf, camel, sweet_pepper, spider, caterpillar, trout, turtle, dolphin, shrew, boy, crocodile, lamp, bee, tractor, couch, train, bridge, road, telephone, mouse, otter, skyscraper, television, bowl, chimpanzee, castle, flatfish, can, woman, whale, baby, forest, possum, kangaroo, squirrel, house, pear, snail, rocket, sea, pine_tree, seal, clock
