Stable Diffusion XL v0.9 - an early leak that got users excited

In the thriving world of AI image generators, patience is apparently an elusive virtue. Enthusiasts of Stable Diffusion - arguably the most popular open-source image generator on the Internet - are bypassing the wait for the official release of its latest version, Stable Diffusion XL v0.9, which leaked online on Thursday.

A version intended for cloud testing and not yet intended by the creator of Stability AI for public release has hit the public sphere and has begun to spread rapidly. Joe Penna, a member of Stability AI's creator, had a mixed opinion.

"Am I proud of you or... the opposite of proud?" - he pondered in a thread on Subreddit dedicated to the leak.

Meanwhile, his Stability AI colleague Alex Goodwin reported on Reddit that the team really wanted to implement a model before launch that could run on A1111, Stable Diffusion's user-favorite GUI. Now, apparently, that plan will have to be rushed.

"We had hoped to have time to implement before launch," Goodwin wrote, "but now apparently we'll have to hurry."

Stable Diffusion XL represents the pinnacle in the evolution of open source image generators. It is designed to compete with its predecessors and analogs, including the famous MidJourney.

With the ability to generate MidJourney-quality images, the new release of Stable Diffusion has quickly carved out a niche for itself. It retains its open source ethos while offering benefits such as plugin compatibility, privacy assurance, and hundreds of different models.

Such features may appeal to those looking for high-quality graphics without shelling out about $100 a year for MidJourney. In addition, fewer complex prompts are required to produce quality results.

However, there is one caveat to the secret acquisition of Stable Diffusion XL v0.9. The torrent takes up a whopping 91.3 GB of space, although the base model and the tweaker are sufficient for it to work. It also doesn't work well with common interfaces, and the best option for testing it at the moment is the Comfy UI, a lightweight but less intuitive node-based interface.

Stability AI, for its part, is playing catch-up. In an attempt to regain control of the situation, the company has shut down Stable Diffusion XL v0.9 on the Huggingface AI community. Unfortunately for Stability AI, the model is already decentralized via torrent, and so stopping its spread will be difficult.

This new model has a number of drawbacks. Many of the plug-ins that were the key to the success of Stable Diffusion 1.5, such as LoRAs, Lycoris, Lohas, Controlnet and others, are incompatible with this model. In addition, once an image has been acquired using the base model, there is a need for refinement to obtain an optimal final image.

Nevertheless, these glitches have not dampened enthusiasm. The visual quality of the model - 1024x1024 resolution compared to 512x512 in version 1.5 - and the aesthetic quality of the images produced by the XL model are already generating enthusiastic user feedback.

With such an active community, it's only a matter of time before all eyes turn to this new era in generative art, and it seems that the chaos surrounding the leak has unwittingly caused a creative frenzy among users.