Model-J: MAE Model (model_idx_0347)
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 | 0.0001 |
| LR Scheduler | constant_with_warmup |
| Epochs | 7 |
| Max Train Steps | 2331 |
| Batch Size | 64 |
| Weight Decay | 0.005 |
| Seed | 347 |
| Random Crop | True |
| Random Flip | True |
Performance
| Metric | Value |
|---|
| Train Accuracy | 0.9509 |
| Val Accuracy | 0.8611 |
| Test Accuracy | 0.8638 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
bridge, dolphin, chimpanzee, spider, lamp, girl, crab, kangaroo, camel, bottle, lizard, baby, castle, raccoon, tank, trout, rabbit, beetle, flatfish, fox, leopard, orchid, mushroom, turtle, whale, bowl, tiger, cockroach, pine_tree, woman, crocodile, lawn_mower, table, shark, aquarium_fish, motorcycle, snail, shrew, plain, pear, rose, telephone, sea, orange, possum, wardrobe, streetcar, poppy, sweet_pepper, lion
