Stability AI’s Post

We’re delighted to announce that Stability AI has secured significant new investment from a world-class investor group including top-tier institutions such as Greycroft, Coatue, Sound Ventures, Lightspeed, O'Shaughnessy Ventures and renowned operators and tech executives Sean Parker, Eric Schmidt, Robert Nelsen and Prem Akkaraju. Prem, a highly experienced media, entertainment and technology executive, will also join as our new CEO. This is a major milestone for the business as it accelerates growth and continues to deliver world-class generative AI products into the market for consumer and enterprise use. Alongside our new CEO, we are pleased to welcome a seasoned roster of investors and new Board members supporting the business including entrepreneur, philanthropist and former President of Facebook, Sean Parker as Chairman. This is a pivotal moment for the Stability AI mission, for our team, and for the sector as a whole. We’re just getting started. For more details, visit the Stability AI website: https://lnkd.in/ghTjZ_Fh

We’re excited to share that our Head of Public Policy, Ben Brooks, will be hosting at The Linux Foundation‘s AI Dev Conference in Paris this Thursday to explore the impact of future AI regulation on open innovation. Join in as we explore how global AI reform impacts open research, development, and collaboration – and what the community can do about it. Learn more and sign up here: https://sched.co/1d20L

Today, we’re thrilled to announce the open weights for Stable Diffusion 3 Medium, the latest and most advanced text-to-image AI model in our Stable Diffusion 3 series! This new release represents a major milestone in the evolution of generative AI and continues our commitment to democratising this powerful technology. SD3 Medium is a 2 billion parameter SD3 model that offers some notable features: 📷 Overall Quality and Photorealism: Delivers images with exceptional detail, color, and lighting, enabling photorealistic outputs as well as high-quality outputs in flexible styles. Success in addressing common pitfalls of other models, such as realism in hands and faces, is achieved through innovations such as the 16-channel VAE. 🖼️ Prompt Understanding: Comprehends long and complex prompts involving spatial reasoning, compositional elements, actions, and styles. By utilizing all three text encoders or a combination, users can trade off performance for efficiency. ✍️Typography: Achieves unprecedented text quality with fewer errors in spelling, kerning, letter forming, and spacing by leveraging our Diffusion Transformer architecture. 🖥️Resource-efficient: Ideal for running on standard consumer GPUs without performance degradation, thanks to its low VRAM footprint. 🧲 Fine-Tuning: Capable of absorbing nuanced details from small datasets, making it perfect for customisation. 🎉 Learn more and get started here: https://lnkd.in/grXfDWgC

Introducing Stable Video on Stable Assistant – your new tool for turning concepts into captivating cinematics. 📹 Whether you’re wanting to create stunning videos from text or images, or bring your conceptual creative vision to life, Stable Video does it all. Visit https://lnkd.in/eDDx4rmG to learn more.

We’re excited to announce Stable Audio Open, an open source model optimised for generating short audio samples, sound effects and production elements using text prompts. This release marks a key milestone as we further open portions of our generative audio capabilities to empower sound designers, musicians and creative communities. Trained on data from Freesound and Free Music Archive (FMA), Stable Audio Open provides free model weights to download on Hugging Face, documentation & code examples, and local fine-tuning on your custom audio data. Key features include: 🎵Generating up to 47 seconds of high-quality audio clips. 🥁Creating drum beats, instrument riffs, foley and ambient sounds. 🎧Enabling variations and style transfer of uploaded samples. Learn more and get started here: https://lnkd.in/gSATxY8m #GenAI #AudioTech #SoundDesign #MusicProduction #MusicTech #AIforCreatives #StabilityAI #StableAudioOpen #OpenSource

What an incredible moment at COMPUTEX TAIPEI! 🎉 Thank you to AMD for including Stability AI in the opening keynote speech alongside all the other incredible partners who are pushing the envelope with the next generation of high-performance AI, PC, and data centre solutions. This was the moment co-CEO Christian Laforte shared the announcement of the Stable Diffusion weights release on June 12th. We can't wait to share our world class model with you next week and see all your marvellous creations. #StableDiffusion3 #StableDiffusion3Medium #StabilityAI #AMD

"We aim to 10x the speed of video game creation and filmmaking". SD3 Medium provides unprecedented visual quality, running super fast on DMI300. It supports simple prompts , ensuring access to all levels of AI model expertise, whilst also understanding complex prompts, bringing all elements of the request together rapidly. SD3 excels at all artistic styles and photorealism and has been trained to produce realistic complex imagery for items such as fingers, guitar strings and other intricately detailed items. Watch the full announcement on the SD3 weight release at COMPUTEX TAIPEI by our co-CEO Christian Laforte earlier this morning. https://lnkd.in/gbvPwGED

The “weight” is nearly over! Today, at COMPUTEX TAIPEI, our Co-CEO, Christian Laforte, officially announced the open release date of Stable Diffusion 3 Medium for June 12th. 🔗 Sign up to the waitlist to be the first to know when the model releases: https://lnkd.in/gij8ZEFt 

Have you seen our Stable Diffusion 3 research paper? 📝 When researching Stable Diffusion 3 (SD3) against other state of the art, text to image generation systems such as DALL·E 3, Midjourney v6 and Ideogram, we found: 1. Stable Diffusion 3 outperforms all models in typography and prompt adherence, based on human preference evaluations. 2. Our new Multimodal Diffusion Transformer (MMDiT) architecture uses separate sets of weights for image and language representations, which improves text understanding and spelling capabilities, even when compared to previous versions of Stable Diffusion. There will be multiple variations of Stable Diffusion 3 available during release stages, ranging from 800m to 8B parameter models to further eliminate hardware barriers. The research paper can be found on our website on the link below. https://lnkd.in/e_8tf9xz #StableDiffusion3 #SD3 #texttoimage #AIimages #AIResearch