---
tags:
- image-classification
- pytorch
- huggingpics
metrics:
- accuracy

model-index:
- name: ViT
  results:
  - task:
      name: Image Classification
      type: image-classification
    metrics:
      - name: Accuracy
        type: accuracy
        value: 0.9299362897872925
---

# ViT

Created a Simple Image Classifier using ViT Transformers, custom trained on blur-dataset with an acc of about 93%.
The Dataset link : https://www.kaggle.com/datasets/kwentar/blur-dataset

## Architecture

![architecture](images/arch.gif)

## Example Images


#### defocused_blurred

![defocused_blurred](images/defocused_blurred.JPG)
#### motion_blurred

![motion_blurred](images/motion_blurred.JPG)

#### sharp

![sharp](images/sharp.JPG)