https://github.com/rafaykhattak/captionify

With Captionify, users can upload an image or enter an image URL to generate a descriptive caption that accurately describes the contents of the image.
https://github.com/rafaykhattak/captionify

deep-neural-networks encoder-decoder-model gpt-2 transfer-learning transformer vision-transformer

Last synced: 29 days ago
JSON representation

With Captionify, users can upload an image or enter an image URL to generate a descriptive caption that accurately describes the contents of the image.

• Host: GitHub
• URL: https://github.com/rafaykhattak/captionify
• Owner: RafayKhattak
• Created: 2023-05-04T12:07:07.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2023-05-04T14:17:11.000Z (almost 2 years ago)
• Last Synced: 2024-11-13T02:32:20.734Z (3 months ago)
• Topics: deep-neural-networks, encoder-decoder-model, gpt-2, transfer-learning, transformer, vision-transformer
• Language: Python
• Homepage: https://rafaykhattak-captionify-app-btvva7.streamlit.app/
• Size: 16.6 KB
• Stars: 5
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Captionify

Captionify is a web application that generates a descriptive caption for an image using an encoder-decoder architecture. The application uses a pre-trained Transformer-based vision model (ViT) as an encoder and a pre-trained language model (GPT2) as a decoder to generate highly accurate captions for uploaded images or image URLs.

![Screenshot (437)](https://user-images.githubusercontent.com/90026724/236234966-71693df3-f3a7-45f2-8a90-109c85315d6e.png)

## Usage

To use Captionify, simply upload an image or enter an image URL on the web interface. The tool will then use the pre-trained models to generate a descriptive caption that accurately describes the contents of the image.

## Getting Started

To install Captionify, simply clone this repository and install the necessary dependencies using pip:

```

git clone https://github.com//.git

cd 

```

Install the required dependencies using the following command:

```

pip install -r requirements.txt

```

Then run the app.py file using the following command:

```

streamlit run app.py

```

This will launch the application on your local machine. You can then upload an image or enter an image URL to generate a descriptive caption.

## Architecture

Captionify uses an encoder-decoder architecture to generate captions for images. The encoder is a pre-trained Transformer-based vision model (ViT) that encodes the input image into a sequence of feature vectors. The decoder is a pre-trained language model (GPT2) that generates a descriptive caption for the image based on the encoded features.

![vit_architecture](https://user-images.githubusercontent.com/90026724/236233200-745dae6a-569f-4558-9a12-3a56b0b8a872.jpg)

## Dependencies

- streamlit

- requests

- Pillow

- transformers

- torch

## References

- This project is based on the Encoder-Decoder architecture and uses pre-trained models from the Hugging Face Transformers library.

- The application was developed using Streamlit, an open-source app framework for Machine Learning and Data Science projects.