What is Stable Diffusion 2.0?

Stability AI Just announced Stable Diffusion 2.0!

Hey Guys,

So when Stability.AI announced Stable Diffusion in August, 2022, it was really a transformative time. Since then the English startup has raised significantly more funding.

You can read their announcement here, now that GPT-4 should be announced soon, things are look promising in the Generative A.I. future.

Are we in a Cambrian Explosion of software 3.0? You tell me. I’ve spent some of the last few weeks thinking about it and the news doesn’t stop coming. Even Open A.I. is investing in the sector.

I will try to highlight some of what stands out for me. It’s the weekend for me, so I’m not going to go into a heck of a lot of detail.

Stable Difusion 2.0 was announced on November 23rd, 2022.

Stability AI is building open AI tools that will let us reach our potential. Designing and implementing solutions using collective intelligence and augmented technology. Their headcount according to LinkedIn has grown nearly 250% in the past six months.

Many new features in v2:
• Base 512x512 and 768x768 models trained from scratch with new OpenCLIP text encoder
• X4 upscaling text-guided diffusion model
• New “Depth2Image” functionality

Blog: https://stability.ai/blog/stable-diffusion-v2-release

Github: https://github.com/Stability-AI/stablediffusion

See on Hugging Face: https://huggingface.co/stabilityai/stable-diffusion-2

Generative A.I. startups are going to really start to form between now and 2024, as it’s clear Venture Capital is willing to invest in the trend for LLMs and text-to-xyz tech. The is an upgrade for some aspects of task automation in the knowledge economy. Over the next decade, what will it bring?

What the Company Says

It is our pleasure to announce the open-source release of Stable Diffusion Version 2.

The original Stable Diffusion V1 led by CompVis changed the nature of open source AI models and spawned hundreds of other models and innovations all over the world. It had one of the fastest climbs to 10K Github stars of any software, rocketing through 33K stars in less than two months.

Source: A16z and Github

• See the light blue line on the left? That’s the rate of adoption of the Stable Diffusion trend.

According to Google trends in the “News” category hype about it has been relatively high in the second half of 2022 on a global basis:

If I change the Google trends to all-categories, it’s even more bullish.

Since 2022 was the year of generative AI, users were able to generate from text to anything. The year isn’t even over yet, and Stability AI has announced the open-source release of Stable Diffusion 2.0 on Thursday. This coincided with American Thanksgiving so you may not have heard the news.

A Bit of Background

The dynamic team of Robin Rombach (Stability AI) and Patrick Esser (Runway ML) from the CompVis Group at LMU Munich headed by Prof. Dr. Björn Ommer, led the original Stable Diffusion V1 release.

Support from Eleuter AI and LAION

They built on their prior work of the lab with Latent Diffusion Models and got critical support from LAION and Eleuther AI.

LAION are one of the “decentralized research collectives” I often go on about. They are a non-profit organization with members from all over the world, aiming to make large-scale machine learning models, datasets and related code available to the general public.

You can read more about the original Stable Diffusion V1 release in their earlier blog post. Robin is now leading the effort with Katherine Crowson at Stability AI to create the next generation of media models with our broader team.

Stable Diffusion 2.0 delivers a number of big improvements and features versus the original V1 release, so let’s later take a small dive in and take a look at them.