metadata
base_model: facebook/dino-vitb16
library_name: transformers
pipeline_tag: image-classification
tags:
- probex
- model-j
- weight-space-learning
Model-J: DINO Model (model_idx_0648)
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
🌐 Project | 📃 Paper | 💻 GitHub | 🤗 Dataset
Model Details
| Attribute | Value |
|---|---|
| Subset | DINO |
| Split | test |
| Base Model | facebook/dino-vitb16 |
| Dataset | CIFAR100 (50 classes) |
Training Hyperparameters
| Parameter | Value |
|---|---|
| Learning Rate | 0.0003 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 3 |
| Max Train Steps | 999 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 648 |
| Random Crop | True |
| Random Flip | False |
Performance
| Metric | Value |
|---|---|
| Train Accuracy | 0.4084 |
| Val Accuracy | 0.3683 |
| Test Accuracy | 0.3682 |
Training Categories
The model was fine-tuned on the following 50 CIFAR100 classes:
bicycle, snake, apple, motorcycle, pickup_truck, worm, sea, mouse, keyboard, bear, orange, crocodile, caterpillar, bowl, ray, cup, shrew, rose, rocket, telephone, hamster, aquarium_fish, cockroach, boy, rabbit, man, lobster, clock, wardrobe, squirrel, can, table, crab, lion, beaver, cattle, bottle, shark, elephant, mountain, castle, forest, chair, mushroom, pine_tree, wolf, dinosaur, poppy, oak_tree, porcupine