Integrate state of the art image generation capabilities directly into your robot apps and products.
How to add the Openai Dall-E Image robot skill
- Load the most recent release of ARC (Get ARC).
- Press the Project tab from the top menu bar in ARC.
- Press Add Robot Skill from the button ribbon bar in ARC.
- Choose the Camera category tab.
- Press the Openai Dall-E Image icon to add the robot skill to your project.
Don't have a robot yet?
Follow the Getting Started Guide to build a robot and use the Openai Dall-E Image robot skill.
How to use the Openai Dall-E Image robot skill
Integrate state-of-the-art image generation capabilities directly into your robot apps and products. DALLE 2 is a new AI system that can create realistic images and art from natural language descriptions.
Have your robot programmatically generate images from speech recognition robot skills descriptions. Or, have the image from the camera sent to Dall-e for its AI interpretation. Experience the whacky world of AI image generation with your Synthiam-powered robot or webcam.
How Does It Work?
DALLE 2 has learned the relationship between images and the text used to describe them. It uses a process called "diffusion," which starts with a pattern of random dots and gradually alters that pattern towards an image when it recognizes specific aspects of that image. DALLE 2 can create original, realistic photos and art from a text description. It can combine concepts, attributes, and styles.
Main Window
2) Modify the Camera Image Use an image from the camera device and run it through the AI variation. This will allow Dall-e to make a new version of a similar but different image.
Configuration
1) API Settings Enter the API Key from OpenAI's website here. This value will be saved locally on the current user account and used for any Open AI robot skills. This value will not be distributed with your project.
Control Commands
There are a few ControlCommand() that allow programmatic usage of this robot skill from other robot skills. For example, send the image to be described by the result of a speech recognition request. The Variation option, where the robot's camera image is modified, uses the camera device robot skills, so they are automatically queried for the control command. If no camera device name is specified, the first camera device is used.- GetImageFromText Example: ControlCommand("OpenAI Dall-e Image", "GetImageFromText", "description of the image to generate"); Generate an image based on the provided description.
- GetImageVariation Example #1: ControlCommand("OpenAI Dall-e Image", "GetImageVariation"); Example #2: ControlCommand("OpenAI Dall-e Image", "GetImageVariation", "Camera"); *Note: When no camera device name is specified, the first camera device will be used.
API Key
An account is required on the OpenAI website to use the service. Once an account is created, you will have an API key to use in the robot skill.Visit the OpenAI website and select the API option from the menu.
Press the SIGNUP option to create an account
Once the account is created, press your Username icon to view the drop-down. And select the View API Keys option.
Copy your API key
Paste the API key in the robot skill configuration menu
Haha this is so neat. I used the Open AI Chatbot robot skill and combined the output with this Dall-e robot skill. I asked Davinic to tell me a story about a bird that wanted to be a pro skate boarder. I fed each sentence into Dall-e and here's the story and images I got...
The story begins...
Once upon a time there was a little bird who left the nest to become a professional skateboarder.
The bird was very small, but it was determined to follow its dream
It practiced hard and eventually became the best skateboarder in the world.
Here's another good one about
Once upon a time, there was a fish who dreamed of becoming a champion wrestler
One day, his dream came true when he sprouted legs and began competing in the ring.
The fish quickly rose to the top, becoming the most popular wrestler in the world.
He was loved by fans everywhere, and his matches were always exciting to watch.
One day, the fish finally achieved his goal of becoming the world champion wrestler.
He was the first fish wrestler to ever achieve this feat, and he was very proud of his accomplishment.
Lol, I do like the story about a bird that wanted to be a pro skateboarder! The fish story definitely had some odd pictures, but the irony of the story is that it doesn't matter how tough of a fish you are, if you take the bait you'll still get hooked in the end
I entered: "the future of humanity" and it looks like we'll all be babies with alien feet using a slightly off-color Windows operating system
Updated to version 3
I forgot to say a big "Thank You!" to Synthiam! I really wanted to play will Dall-E images, but I didn't know how to get access to it without being a developer. I felt like I was missing out on a lot of fun Now, I can join in the fun quite easily using my favorite robot software!
Oh, man! We could get all sorts of inspiration from Dall-E for robot builds! We could take the Dall-E concepts, design them in 3D and bring them to life with EZ-Robot parts + ARC. My mind is overwhelmed by the possibilities at the moment!!!
This is one of the pictures I got when I entered "generation 1 transformers" it's a whole new breed of Autobots!
Another picture I received looked like the robots were made of bad origami and tin foil
Thanks, @Dave!! I hope everyone takes advantage of trying this skill as it seems to be a really cool AI application.
I'm still learning how to use it effectively. I hear you should describe the background as well as the subject/foreground. You can also choose the style by writing "oil painting" or "Claude Monet".
Here's the description I used to create a picture I had in my mind.
"2 highly detailed ninja robots with a dark and neon Japanese cityscape behind them on a foggy night"
While it wasn't exactly what I wanted I did appreciate the creative output!
It's a very interesting AI. Seems like a lot of fun and a good way to pass some time away. Kinda gets you thinking.