Model-J: MAE Model (model_idx_0313)
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 | val |
| Base Model | facebook/vit-mae-base |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|
| Learning Rate | 0.0005 |
| LR Scheduler | linear |
| Epochs | 8 |
| Max Train Steps | 2664 |
| Batch Size | 64 |
| Weight Decay | 0.007 |
| Seed | 313 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.8920 |
| Val Accuracy | 0.5536 |
| Test Accuracy | 0.5522 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
couch, mushroom, shark, mouse, flatfish, girl, raccoon, squirrel, trout, streetcar, train, plate, camel, caterpillar, mountain, willow_tree, skunk, baby, bee, shrew, oak_tree, ray, pine_tree, worm, tiger, sea, elephant, clock, wolf, bottle, bed, whale, beetle, aquarium_fish, road, orange, cockroach, cloud, palm_tree, cup, skyscraper, rose, butterfly, turtle, apple, woman, telephone, seal, chair, bowl
