Skip to content

Generate LEGO like looking images with gpt4-vision and DALL·E 3.

Notifications You must be signed in to change notification settings

thomasrosen/lego_vision

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

LEGO Vision

Generate LEGO like looking images with gpt4-vision and DALL·E 3.

This uses gpt4-vision to generate a prompt for DALL·E 3. The resulting image should be in the style of the LEGO movie.

🎈 This is inpsiered by @jh3yy's tweet: https://twitter.com/jh3yy/status/1730596577067122956.

One example

More examples are below.

A family in Berlin

1. add your own API key

You need to add your own OpenAI API key to a .env-file. There is an example.env-file for help with naming.

The .env-file should look like this:

OPENAI_API_KEY="sk-xxxxxx"

2. install dependencies

npm install

3. Edit the script

You need to add the path to your image and maybe some additional info about the photo to the script.

In the gen_image.js file you only need to edit the main-function. (You can of course edit the rest of the file as you like :) )

You can use the image_url if the image is hosted somewhere. Otherwise use the image_path to define a relative path.

const new_image_url = await the_chain({
  image_path: './path_to_the_image.jpg',
  image_url: 'https://url.to/the_image.jpg',
  additional_info: 'The age of the person or other important info.', // leave empty if not needed
})

4. run the script

Running the script can take a few seeconds. It will print out the prompt it send so DALL·E 3 and the final image url to download the image.

node gen_image.js

Examples

A person near some water Two people kissing A model in a castle A selfie in a livingroom A young person in the livingroom

About

Generate LEGO like looking images with gpt4-vision and DALL·E 3.

Topics

Resources

Stars

Watchers

Forks