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_0686)

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 | 0.0003 |
| LR Scheduler | cosine_with_restarts |
| Epochs | 4 |
| Max Train Steps | 1332 |
| Batch Size | 64 |
| Weight Decay | 0.03 |
| Seed | 686 |
| Random Crop | True |
| Random Flip | False |

## Performance

| Metric | Value |
|---|---|
| Train Accuracy | 0.9956 |
| Val Accuracy | 0.9312 |
| Test Accuracy | 0.9332 |

## Training Categories

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

`caterpillar`, `maple_tree`, `bee`, `fox`, `turtle`, `camel`, `bus`, `can`, `shrew`, `bicycle`, `spider`, `skunk`, `crab`, `bowl`, `tank`, `lawn_mower`, `man`, `plain`, `leopard`, `oak_tree`, `boy`, `pickup_truck`, `beetle`, `forest`, `otter`, `bed`, `television`, `house`, `ray`, `tiger`, `rocket`, `mushroom`, `wolf`, `apple`, `sweet_pepper`, `mountain`, `whale`, `skyscraper`, `pear`, `pine_tree`, `table`, `dolphin`, `crocodile`, `tulip`, `snake`, `flatfish`, `aquarium_fish`, `sea`, `worm`, `rose`