Prompt Art

🎨 Create stunning art with AI.
Just describe what you want to draw in words and the AI will paint the picture for you in seconds.
👇 Try it yourself

[Daily AI Papers @papers_daily]

Exciting news! Our iOS app received a major update. For just $0.99/month, you can now filter and sort AI research papers by monthly popularity and published date.

Download and let us know what you think

📱⬇️ https://t.co/dwjNMhwvCH

#AIresearch #iOSapp #researchpapers

[Prompt Art @PromptArtApp]

We regret to inform that we are discontinuing PromptArt iOS and web apps. We couldn't generate enough revenue from subscriptions to cover the server costs.😢

We will stop generating capability on Saturday and keep the apps live till 28th, Feb for you to download your creations.

[Prompt Art @PromptArtApp]

Our site will undergo maintenance for approximately 30 minutes. We apologize for any inconvenience this may cause.

[Prompt Art @PromptArtApp]

Generate avatars of yourself with Prompt Art app by uploading 10+ photos. Apart from 50+ samples, you can generate as many avatars with your own prompts.

It will be available free for subscribed users ($0.99/mo) while we are beta testing.

📱 https://t.co/TsYl6173mF

[Prompt Art @PromptArtApp]

7/7) Explanation of the prompt:

"Street Photography" gives a good angle and adds people and vehicles to the art

"Art by Craig Mullins" make it darker and more art like (in contrast to a 3D rendering)

"Mystery Novel Cover" makes it more dramatic and sometimes add more details

[Prompt Art @PromptArtApp]

1/7) Stunning art of streets from around the world

Generated with https://t.co/TsYl6173mF

#StableDifusion #AIArtwork #workflow

Here are some awesome arts generated:

🧶👇

[Prompt Art @PromptArtApp]

6/7) It works pretty well for fantasy and fiction too. These are from Tatooine (Star Wars) and Shire (LOTR)

👇

[Prompt Art @PromptArtApp]

Instruct-pix2pix Paper: https://t.co/oUjwvHfHji

[Prompt Art @PromptArtApp]

Introduced image editing with instructions to PromptArt App with instruct-pix2pix

Appstore: https://t.co/lFz5Zv9O7C

[Prompt Art @PromptArtApp]

We found two bugs in our (beta) avatar generator.

1. Nothing happens when you tap “Submit to training” if the number of uploaded images is less than 20.
2. The popup to confirm doesn’t show when you tap “Upload Photos” on iPads.

We will fix this and release an update shortly.

[Prompt Art @PromptArtApp]

(2/2) emotional variations (suggestion in Reddit👇)
https://t.co/IMhZeMOTLC

[/MachineLearning @slashML]

Using my wife's photos to finetune an SD 768x768 2.1v model and experimented with iconic movie characters #stablediffusion #gpt3 https://t.co/KJm558kIBZ

[Prompt Art @PromptArtApp]

Our team member @hnipun1 finetuned an SD 768x768 2.1v model using his wife's photos and experimented with iconic movie characters.

750 training steps, 0.000003 learning rate, and around 20 images used

#stablediffusion #stablediffusionart
👇🧵

[Prompt Art @PromptArtApp]

2/2 The generated image is lacking some details, but looks almost like him in real life (except for the crown and dress).

Will post updates when we can generate higher quality images.

[Prompt Art @PromptArtApp]

Experimenting with image to image editing with user images

Original image on left taken from our feed: https://t.co/bpj1NIPbdp

The photo on right is the new generation with our team member @nmasnadithya

We used #stablediffusion 2.1v model and dreambooth training

👇

[Prompt Art @PromptArtApp]

11/11

Here’s an annotated PyTorch implementation of stable diffusion for a more detailed look into it : https://t.co/jbzicS4DZd

[Prompt Art @PromptArtApp]

Want to train stable diffusion with your own images? Here's what you need to know...

🧶👇

[Prompt Art @PromptArtApp]

10/n

Autoencoder: maps images to and from the latent space, reducing resolution by a factor of 8 and the latent space has 4 channels instead of the 3-channel (RGB) image space.