Add Vision Transformer (ViT) Demo to computer_vision Module

[dhruvidave348]

Feature description

This PR adds a fully functional demo of a Vision Transformer (ViT) for image classification using Hugging Face Transformers.

Features included:

• Loads a sample image from the web.
• Uses ViTImageProcessor for preprocessing.
• Performs inference with ViTForImageClassification.
• Prints the predicted label for the image.
• Handles network timeout and correct import order.

Example output:
Predicted label: tabby, tabby cat

This demo can be used as a reference for anyone learning ViT or image classification with Hugging Face Transformers.

Additional Notes:

• Requires torch, transformers, and PIL.
• Can be extended to classify local images easily.

Hacktoberfest 2025: ✅

[hitdepani]

I WOULD LIKE TO WORK ON THIS

[Khansa435]

Hi! I’d like to work on this issue for Hacktoberfest. Could you please assign it to me?

[TASMAYU]

Hey @dhruvidave348
I’d like to work on this issue as part of Hacktoberfest 2025.
I’ve gone through the description , the goal is to add a fully functional demo for Vision Transformer (ViT) using Hugging Face’s transformers library for image classification.

Here’s my approach:

1.Load a sample image (from web or local).
2.Use ViTImageProcessor for preprocessing.
3.Perform inference using ViTForImageClassification.
4.Display the predicted label clearly.

Handle network timeout and import order correctly.
Using it with torch, transformers, and PIL library
I can also add brief comments to make it easy to understand and to extend the demo as well.

Please assign this issue to me
I am starting working on it

[Wvidit]

Can you elaborate a little more, like what model to use for inference and what image inference defaults to?