This is a simple front + back making it easy to use DALL-E models

Another option is to use the gradio demo

If you want to run the backend on google colab, you can run this notebook

You can then use the gh pages front with an url such as https://rom1504.github.io/dalle-service?back=https://XXXX.loca.lt

First follow back (you may choose to use https://ngrok.com/ to expose your locally running backend)

Then put your backend url in https://rom1504.github.io/dalle-service/

You can share an url such as https://rom1504.github.io/dalle-service?back=https://yourbackend.com

If you want to run everything yourself, you can go to back then to front

You can either train your model yourself with DALL-E

Or use a pretrained one from https://github.com/robvanvolt/DALLE-models/tree/main/models/taming_transformer

• laion subset cars a model trained on 10k cars extracted from laion 400m