All Your Tech posts
Posts: 118
Files: 359
Filtering and Upscale Image Saving Now Live at PixelDojo.ai
One of my goals with PixelDojo.ai is to help people use cutting edge AI art tools like upscalers without the usual huge learning curve. As part of this, I recently launched the Community Gallery, a place where people can share their creations along with the settings needed to create a similar image. Now I have added the ability to save and view other users upscaled images. Not only can you see examples of people using the upscaler along with the before and after image, but you also get to see all of the settings used so you can improve your own...
"Generate This Image" Now Live At PixelDojo
A couple weeks back I added aCommunity Galleryto PixelDojo.ai. The idea was simple, we have this amazing community to creative people generating some really cool AI art. Why not have a way to share those creations with each other so we can all learn and benefit? Now, I added the next layer. When you view an image in the gallery, you can now click the "Generate This Image" button. That loads up ALL of the prompt information and parameters in the Stable Diffusion AI Image generator OR in the Stable Diffusion 3 image generator so you have...
New Cutting Edge Image Relighting Tool Now At PixelDojo.ai
Generative AI tool number 13 has been officially added to PixelDojo.ai. This one allows you to upload a source image, then use a prompt to control the lighting and overall mood of the image. From a dimly lit cyberpunk city street, to a sunny European vilage and everything in between.Pair this with our newly released Consistent Characters tool, or the creative upscaler to take your photos and AI generated images to the next level. Try it out here: https://pixeldojo.ai/image-relighting
New State-Of-The-Art Consistent Character generator at PixelDojo.ai
I just finished adding a new Consistent Characters tool to the system. This one is by far the best I have tested, and i'm so happy to be able to add it to the Dojo.Upload a single image, and let the magic happen. The system will generate 4 additional images in carious poses while maintaining the overall look and feel of the original character. It's also fully controllable via prompting, so you can change the clothing, style, environment, etc.Enjoy! Try it out here: https://pixeldojo.ai/consistent-characters
PixelDojo Community Gallery
Pixeldojo now has a new Community Gallery. Here we can share prompts and images we create with the entire world. Consider it a beta release with many more features to come To add your images:1) create an image with The AI Image Generator or Stable Diffusion 32) save the images you love to “my images”3) view My Images and look for the icon in the bottom right of the images. By default this shows a lock because all of your saved images are private by default.4) Click the lock icon and the image changes to public so the world can see it.Comme...
UPDATED: SDXL Local LORA Training Guide: Unlimited AI Images of Yourself
✨ Config Files: https://www.patreon.com/posts/koya-ss-tutorial-95710740✨ Kohya SS Auto Installer: https://www.patreon.com/posts/1-click-for-ss-103541233✨ regularization images: https://www.patreon.com/posts/koya-ss-tutorial-95710740💻My Stable Diffusion PC: https://kit.co/AllYourTech/stable-diffusion-buildThis guide contains everything you need to train your own LORA or Low Rank Adaptation model for Stable Diffusion XL (SDXL) using your home PC. This guide will allow you to train SDXL to generate images of yourself, or anyone else for...
PuLID Character Consistency now live @ PixelDojo
Update time!I just pushed out a new consistent character creator based on the JUST released PuLID (Pure and Lightning ID Customization via Contrastive Alignment) method. It may be difficult to say, but the results speak for themselves. It's VERY fast using a lightning model and only 4 steps to generate an image, BUT it is also high quality. The implementation at https://PixelDojo.ai returns 4 images for every generation.and.... since we now have two completely different methods of generating consistent characters, I renamed the old Consist...
1 Click Installer for Kohya SS (LoRA Training GUI)
Hello everyone! Bom Dia! (I moved to Portugal, that means Good Day).I've created a 1-click auto installer for KOHYA SS, a tool you have seen me use to train LoRA models for Stable Diffusion and SDXL. This batch file automates the entire install process with just 1 CLICK!Here's how it works:1. Download the batch file from down below and put it in an empty folder somewhere on your computer.2. Double click the batch file to run it. The file will automatically:- Install Python and other dependencies IF you don't have it installed alrea...
Creative Upscaler Updates @ PixelDojo
Alright, time for more updates over at https://pixeldojo.ai. I added a new high dynamic range slider that adds more vibrant colors to washed out images, a sharpen feature that runs a sharpener after the image is upscaled (defaults to off), increased the max upscale to 8X (takes a LONG time), and finally increased the range for the resemblance from 0-1 to 0-3. For those of you doing interior renderings (cough, cough), setting a low creativity and high resemblance should retain the original textures better. If that works i'll add some default...
The PixelDojo Affiliate Program is here
Earn a 40% commission for every new signup you send over. Join here! pixeldojo.getrewardful.com/signupHere is what people are saying: "This is unbelievably good. Pixel Dojo's Upscalers are simply the best I've ever seen." - Silvio S.
Image to prompt is now live at PixelDojo
Image to prompt is now live at PixelDojo. This allows you to upload an image and return a prompt. Useful to give you ideas when using upscalers and image to image models. If we find that this is useful I will add this as an automatic step when upscaling. Select an option to automagically create a prompt, and this will happen in the background.https://pixeldojo.ai
New Multidiffusion Upscaler @ PixelDojo
An all-new Multidiffusion Upscaler has been added to the suite of AI tools @ https://pixeldojo.ai This new upscaler creates sharp, detailed images without additional tuning and tweaking which is rare with these tools. You now have access to all of these with one subscription: Unlimited AI generations Access to the latest Stable Diffusion XL modelsStable Diffusion 3 and SD3 TurboCreative Upscaler Multidiffusion UpscalerCreate Consistent CharactersAccess to uncensored modelsPriority supportStyle TransferAI Face EnhancerImage to Image and more...
Face Enhancer Beta @PixelDojo
We've all had it happen, we get an overbaked image back from stable diffusion, or that typical "AI Look" in an otherwise great image. This new feature is designed to fix these issues by using a transformer-based prediction network that takes a low-quality face and discovers a close approximate target, even when the source image is highly degraded. The result is higher-quality images with the ability to control the tradeoff between quality and fidelity/accuracy.This new feature will go live tonight at https://pixeldojo.ai/face-enhan...
PixelDojo.ai price increase in 24 hours!
PixelDojo.ai price increase in 24 hours! Tomorrow the price of a monthly subscription will increase for NEW subscribers. We will be going from $9.95 per month to $19.95 per month in order to cover GPU costs. This is a great opportunity for you to subscribe BEFORE the price increase: https://pixeldojo.ai Don't forget you get access to:Stable Diffusion 3 Creative upscalerStable Diffusion 3 TurboStable Diffusion 3 Image-To-ImageStyle TransferStable Diffusion XLConsistent CharactersJuggernaut XLPlayground XLKandinskyDreamshaper XLRealVis XLwith...
How to use the Stable Diffusion 3 API
Step 1: Create an accountCreate an account on Stability AI Developer platform.Step 2: Add creditsA new account should come with 23 free credits when you sign up. This is enough for about 3 images. Add more credits on the billing page. The minimum is $10 ($12.30 for europeans)Each image generation consumes credits. See the pricing pagefor credit usage for each image.Step 3: Create an API keyCreate an API key on your account’s API keys page. You can create a new one or use an existing one.Copy the API key.Step 4: Generate imagesOpen this Colab...
Stable Diffusion 3 @ PixelDojo
You can either pay $10 to buy credits to test Stable Diffusion 3 at StabilityAI (12.30 if you are in the EU), or Pay 9.95 for a subscription to https://PixelDojo.ai which features 6 cutting-edge Generative AI Tools including Stable Diffusion 3 and SD3 Image-to-Image.
Style Transfer is now live @ PixelDojo.ai
Style Transfer is now live at https://PixelDojo.ai . Supply an image, and add a style image to instantly reimagine a new image that combines the look of the original with completely new aesthetic and style.
Stable Diffusion 3 Image-To-Image live at PixelDojo.ai
Stable Diffusion 3 Image-To-Image generation is now live at https://PixelDojo.aiProvide a source image and a prompt and see how you can creatively modify your images.
PixelDojo adds aspect ration controls for Stable Diffusion 3
PixelDojo Stable Diffusion 3 has been updated with new aspect ratio controls. Check it out at https://PixelDojo.AIImage-To-Image generation should be live within the hour :)
LlaMA 3 was released today
LlaMA 3 was released today, and early benchmarks show that it is VERY impressive for an open source model. Should I add it to PixelDojo?
Stable Diffusion 3 Hands On Review
Stability AI released Stable Diffusion 3 and Stable Diffusion 3 Turbo via API, so I got to work building out a user interface over at PixelDojo.AI where you can use both models to generate your own images. How good are these models? Were the press photos all cherry-picked, or is the text adherence and prompt adherence as good as they claim? Check out the hands-on review and judge for yourself!
Stable Diffusion 3 is now live on PixelDojo.AI
A few hours ago Stability AI made Stable Diffusion 3 and Stable Diffusion 3 Turbo available via their API. You still can't download and run it locally, BUT I was able to code up a UI for it that you can access if you are a subscriber over at https://pixeldojo.ai . It's pretty expensive to run compared to the older models, but I couldn't pass up adding it to the site for people to play around with.Here is the announcement from Stability: https://stability.ai/news/stable-diffusion-3-apiKey TakeawaysStable Diffusion 3 and Stable Diffus...
Boston Dynamics Atlas humanoid robot goes electric
Boston Dynamics has introduced an all-electric version of its Atlas humanoid robot. This new iteration marks a significant departure from the hydraulic systems of its predecessors, offering a quieter operation with the whirring of electric motors. The design of the robot has been significantly overhauled, featuring a more streamlined and cable-free appearance, resembling a softer, more cartoonish figure compared to the previous models.The electric Atlas demonstrates improved fluidity in movement, though still with a hint of jerkiness, reminisce...
Suno AI Competitor UDIO - Early Access
Hey everyone, I took some time to check out the new Suno AI Generative Music competitor Udio. The video goes up tomorrow, but I wanted to give everyone here early access. Enjoy!
New Creative AI Upscaler @ PixelDojo.ai
I spent the last few weeks working on a new creative upscaler to replace the "old" upscaler at PixelDojo.ai. I built this after studying Magnific. This new upscaler not only increases the resolution of the original image up to 4X, it also runs in about half the time and adds an incredible level of detail to images. It uses MultiDiffusion, ControlNet tiles and detail LoRas. It's designed to allow you to easily control the amount of hallucination, resemblance and creativity with the advanced settings so you can get just the right amou...
Suno AI V3 Update - How is it?
Suno AI dropped version 3 of their generative AI music service. I've discussed it at length before showing people how to make full length songs, and how to use various tags and such to control the output, but how is the new version really? Check out the video and take a look!
StableSwarmUI released
Stability AI has released an exciting new project called “Stable Swarm UI.” This revolutionary tool allows you to utilize GPUs not only on your local network but also in the cloud, unlocking tremendous potential to scale and speed up complex workflows. The UI is Alpha, and it definitely needs some work, but here are the details if you want to check it out yourself. You may be able to use Google Colab to run the UI, although I have had mixed results: Colab link if you want to try Swarm: https://colab.research.google.com/github/Stability-AI/S...
AI News 4/14
OpenAI has terminated researchers Leopold Aschenbrenner and Pavel Izmailov for allegedly leaking information outside of the company. The actual information leaked is currently unknown.Leopold Aschenbrenner was part of OpenAI's "superalignment" team and an ally of chief scientist Ilya Sutskever. Pavel Izmailov also spent time on the AI safety team. In case you didn't know, Sutskever was one of the primary figures in the 2023 ousting of Sam AltmanAs reported by Bloomberg, Apple is gearing up to overhaul its entire lineup of Macs w...
New Creative AI Upscaler @ PixelDojo.ai
I spent the last few days working on a new creative upscaler to replace the "old" upscaler at PixelDojo.ai. I built this after studying Magnific over the past few months. This new upscaler not only increases the resolution of the original image up to 4X, it also runs in about half the time and adds an incredible level of detail to images. It uses MultiDiffusion, ControlNet tiles and detail LoRas. It's designed to allow you to easily control the amount of hallucination, resemblance and creativity with the advanced settings so you ca...
AI News April 7th 2024
Researchers from Princeton NLP developed SWE-agent, an open-source project that turns GPT-4 into an AI software engineer. The agent achieved accuracy similar to Devin and resolved 12.29% of issues autonomously.https://twitter.com/_carlosejimenez/status/1775157915789463935OpenAI made a big upgrade to DALL-E 3, now allowing users to edit images directly in ChatGPT. Users can edit images directly in the chat across the web, iOS, and Android apps by selecting an area of the image and prompting changes.https://twitter.com/OpenAI/status/1775569161759...
