Are you ready to drive more awareness to your brand? Consider becoming a sponsor of the AI Impact Tour. Learn more about opportunities here,
Generative AI art has quickly emerged as one of the most interesting and popular applications of the new technology, with models like Stable Diffusion and MidJourney boasting millions of users, with OpenAI’s own DALL-E 3 image generation model bundled directly into The steps to do are not mentioned. Its popular ChatGPT service launched this fall. By simply typing a description and waiting a few moments, users can see an image from their imagination presented on the screen by an AI algorithm trained to do exactly that.
Still, the fact that the user has to wait those “few short moments” for the AI to generate their image, which is anywhere between a second or two to a minute, is our Not ideal for the fast-paced, instant gratification modern world.
That’s why this week, the online AI arts community is collectively discussing a new machine learning technique – LCM-LORA, short for “Latent Consistency Model – Low-Rank Optimization”, developed by the Institute for Developed by researchers from Interdisciplinary Information Sciences (IIIS). Tsinghua University in China and the AI code sharing platform HuggingFace, and described in a paper published on the pre-review open access research site arXiv.org – finally brings generative AI art creation to real-time.
What does this mean in practical terms? Well, check out some videos shared by AI artists on X and LinkedIn below, and you’ll get an idea.
AI Impact Tour
Join the enterprise AI community at VentureBeat’s AI Impact Tour coming to a city near you!
So, is Generative AI here in real time?
Only 4 days have passed and nothing will be the same. All the design tools, all the work processes, everything is going to change.
, Here’s everything you need to know about LCM-LoRa??pic.twitter.com/eK3uoxXBlC
– Javi Lopez ⛩️ (@javiopen) 14 November 2023
I’ve only had access for a short period of time, but this new real-time LCM tool is having a pretty intense effect on my dopamine supply right now.
A whole new era of generative AI is about to begin. pic.twitter.com/M701YoWBpp
– Surea.i (@sureailabs) 14 November 2023
Essentially, thanks to the LCM-LoRA technology, users can now move their cursor or paint simple, almost stick-figure like pictures or apply a few shapes along with descriptive text, and AI art creation applications like Krea.AI and Fal can. The AI will automatically render different, new, generated art instantly, even swapping out imagery in fractions of a second as the user moves their shapes or draws simple lines on their digital canvas. .
You can try it for yourself (if it keeps up with increased usage) here on Fal.AI.
The technology works not only for flat, 2D images, but also for 3D assets, meaning artists can theoretically create instant content for use in mixed reality (AR/VR/XR), computer and video games, and other experiences. Can create immersive environments. Theoretically, they could also be used in films, drastically reducing production costs.
“Everything is going to change,” a startup founder and former Google AI engineer commented on LinkedIn about LCM-LoRA, a sentiment echoed by many in the AI arts community.
“A whole new era of generative AI is about to begin,” Another user commented on,
Ethan Molick, a professor at the University of Pennsylvania’s Wharton School of Business who is one of the most active and vocal influencers and proponents of generic AI, said that thanks to the technology, “we’re going to see a lot of new user experiences very soon. ”
What is LCM-LoRa and how does it work?
Early demos of LCM-LoRa integration in apps are undeniably fascinating and suggest to this writer of VentureBeat/AI artist that this is a new watershed moment for generative AI in the visual arts.
But what is the technological advancement at the core of LCM-LoRa and can it spread across all apps and different uses, as early users indicate?
According to the paper describing the technology published by researchers at IIIS Tsinghua University and HuggingFace, LCM-LoRa is ultimately a “universal training-free acceleration module that can be directly plugged into various stationary diffusion fine-tuned models, or SD LoRa.” “
That’s a mouthful for anyone outside the machine learning community, but to decode it in more common English, it’s essentially an algorithm that uses the popular open-source Stable Diffusion AI to transform text or source imagery into new AI. Speeds up the process of converting generated artwork. The model, and its streamlined, or transformed forms.
LCM-LORA does this by reducing the number of “required sampling steps,” that is, the processes the AI model has to go through to transform the source text or image – whether it is a description or a stick figure – into a high-quality, higher -Detailed image based on learning static diffusion model from millions of images.
This means that LCM-LORA allows static propagation models to work faster with fewer computational resources, so they do not need to take up as much working memory or cycles on an individual’s computer. This is what enables them to deliver amazing results in real time.
The fact that it is “universal” means that it can be plugged into a wide variety of apps that rely on stable diffusion or its variants to generate imagery. Whether this can be extended beyond stable diffusion to proprietary models like OpenAI’s DALL-E 3 or MidJourney remains to be seen.
We’ve contacted one of the authors of the LCM-LoRa paper and will update this piece with more information when they respond.
VentureBeat’s mission To become a digital town square for technology decision makers to gain knowledge about transformative enterprise technology and transactions. Find our briefing.