README.md

---
base_model: google/vit-base-patch16-224
library_name: transformers
pipeline_tag: image-classification
tags:
  - probex
  - model-j
  - weight-space-learning
---

# Model-J: SupViT Model (model_idx_0730)

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

<p align="center">
    🌐 <a href="https://horwitz.ai/probex" target="_blank">Project</a> | 📃 <a href="https://arxiv.org/abs/2410.13569" target="_blank">Paper</a> | 💻 <a href="https://github.com/eliahuhorwitz/ProbeX" target="_blank">GitHub</a> | 🤗 <a href="https://huggingface.co/ProbeX" target="_blank">Dataset</a>
</p>

![ProbeX](https://raw.githubusercontent.com/eliahuhorwitz/ProbeX/main/imgs/poster.png)

## Model Details

| Attribute | Value |
|---|---|
| **Subset** | SupViT |
| **Split** | train |
| **Base Model** | `google/vit-base-patch16-224` |
| **Dataset** | CIFAR100 (50 classes) |

## Training Hyperparameters

| Parameter | Value |
|---|---|
| Learning Rate | 3e-05 |
| LR Scheduler | constant |
| Epochs | 9 |
| Max Train Steps | 2997 |
| Batch Size | 64 |
| Weight Decay | 0.009 |
| Seed | 730 |
| Random Crop | True |
| Random Flip | False |

## Performance

| Metric | Value |
|---|---|
| Train Accuracy | 0.9976 |
| Val Accuracy | 0.9307 |
| Test Accuracy | 0.9386 |

## Training Categories

The model was fine-tuned on the following 50 CIFAR100 classes:

`bee`, `television`, `castle`, `rose`, `sunflower`, `caterpillar`, `mouse`, `otter`, `willow_tree`, `raccoon`, `skunk`, `table`, `crab`, `apple`, `bowl`, `boy`, `kangaroo`, `porcupine`, `beetle`, `flatfish`, `dinosaur`, `squirrel`, `beaver`, `elephant`, `tiger`, `lawn_mower`, `wolf`, `tulip`, `sea`, `snail`, `clock`, `poppy`, `tank`, `butterfly`, `pine_tree`, `bed`, `baby`, `bus`, `ray`, `train`, `aquarium_fish`, `lamp`, `wardrobe`, `girl`, `plate`, `hamster`, `whale`, `fox`, `rabbit`, `oak_tree`