Stable Diffusion is probably the most commonly used image generation model.
Since it is open source, various communities and groups have been fine-tuning it and using it a lot, and it is easy to use with DreamBooth, etc., and modules like Controlnet have been actively developed...
There are also many cases of starting a business by further expanding the model based on Stable Diffusion ( such as Lascaux and Flask in Korea ).
Of course, since it is open source, there are cases where distorted(?), raw desires are projected, but I personally think that this is also a trend.
Stability AI announces new version of Stable Diffusion text-image model SDXL 0.9
SDXL 0.9 produces a vastly improved image and composition detail over previous models.
The model is currently accessible via ClipDrop, with an API coming soon
SDXL 0.9 significantly advances creative use cases for generative AI imagery
SDXL 0.9 is available for Clipdrop, Stability AI API and DreamStudio customers, with a full open release of SDXL 1.0 scheduled for mid-July.
SDXL provides the ability to create hyper-realistic artwork for film, television, music, and educational videos, and drives advancements in design and industry.
The SDXL series goes beyond basic text prompting to provide a variety of features such as inter-image prompting, inpainting, and outpainting.
The main improvement in SDXL 0.9 is the significant increase in the number of parameters, with a 3.5B parameter base model and a 6.6B parameter model ensemble pipeline.
Since the beta launch of SDXL, the Discord community has had over 7,000 users create over 700,000 images, averaging over 20,000 images per day.
SDXL 0.9 runs on two CLIP models using OpenCLIP ViT-G/14 to produce deeper, higher resolution images of 1024x1024.
SDXL 0.9 can only be run in environments with a GPU, and requires Windows 10 or 11 or a Linux operating system, 16 GB of RAM, an Nvidia GeForce RTX 20 graphics card (8 GB VRAM or higher), or a compatible AMD card (16 GB VRAM).
SDXL 0.9 is provided for research purposes and is available for a limited time only to gather feedback and fully refine the model before general public release.
📚 Welcome to Haebom's archives. --- I post articles related to IT 💻, economy 💰, and humanities 🎭. If you are curious about my thoughts, perspectives or interests, please subscribe. Would you like to be notified when new articles are posted? 🔔 Yes, that means subscribe. haebom@kakao.com
