https://github.com/arjunpatel7/perfect-prompt

An approach to creating the perfect prompt for any image generation task.
https://github.com/arjunpatel7/perfect-prompt

cohere nlp prompt stable-diffusion streamlit text-generation

Last synced: about 1 month ago
JSON representation

An approach to creating the perfect prompt for any image generation task.

• Host: GitHub
• URL: https://github.com/arjunpatel7/perfect-prompt
• Owner: arjunpatel7
• License: mit
• Created: 2022-09-30T23:02:19.000Z (about 2 years ago)
• Default Branch: main
• Last Pushed: 2022-12-16T16:50:02.000Z (almost 2 years ago)
• Last Synced: 2024-10-06T09:47:10.736Z (about 1 month ago)
• Topics: cohere, nlp, prompt, stable-diffusion, streamlit, text-generation
• Language: Jupyter Notebook
• Homepage:
• Size: 25.9 MB
• Stars: 29
• Watchers: 1
• Forks: 9
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# :art: Perfect Prompt

An approach to creating the perfect prompt for any image generation task. Type in a prompt, and

Perfect Prompt will match it to one of six art styles (cyberpunk, steampunk, water colors, cottagecore, landscapes,

and architecture) and provide generated variations of your prompt along these lines, in

under a minute!

Want to try it for yourself? Follow the instructions 

below or [click here](https://arjunpatel7-perfect-prompt-app-mckes9.streamlitapp.com/

) for the deployed version!

![demo gif](https://github.com/arjunpatel7/perfect-prompt/blob/main/perfect_prompt_shorter.gif)

## Requirements

* cohere

* streamlit

* replicate (for image generation)

* pandas

## Installation

First, secure API keys for both Replicate and Cohere,

and save them to Streamlit secrets management.

Clone this repo, and run the following command:

```bash

streamlit run app.py

```

## Author

This app was made by Arjun Patel, a data scientist with experience

applying deep learning to audio and text data. 

Connect with me on [Linkedin](https://www.linkedin.com/in/arjunkirtipatel/)!

## Acknowledgements

This app was made during the Cohere AI Hackathon, in which

Cohere and Lablab.ai offered access to Cohere Classify and Generate

APIs for novel uses. Thank you to both groups for their assistance and mentorship.

This app makes use of the Lexica API, which is generously made freely accessible

by the folks over at [Lexica.art](https://lexica.art). Finally, thanks to Replicate for an easy to use api for Stable Diffusion access.