Stable Diffusion is a text-to-image and image-to-image generation model based on Latent Diffusion Models, which can generate high-quality, high-resolution, and highly realistic images based on any text or image input.

eCloudrover provides an out-of-the-box AIGC solution for text-to-image and image-to-image generation, which can help customers quickly generate artistic materials such as pictures and posters.

The core of this solution is an AMI based on Deep Learning AMI GPU PyTorch 1.13.1 (Ubuntu 20.04), which is installed with Stable Diffusion WebUI and integrated with Stable Diffusion 1.5 and ControlNet models.

In this AMI, we also deploy an Nginx application to forward requests from port 80 to port 7860 of Stable Diffusion WebUI. Both Nginx and SD-WebUI are configured to start automatically at boot. When you first start an EC2 instance from the AMI, you may need to wait for 10-15 minutes for Stable Diffusion to load and compile models.

With Stable Diffusion WebUI, you can easily perform text-to-image and image-to-image operations on the web page. By installing WebUI plugins, you can also achieve a higher degree of customization, such as Fine-Tune the model through Dreambooth, import third-party VAE/Lora models, etc.

For the convenience of Chinese speakers, we have installed a Chinese plugin for WebUI by default, which can be disabled on the WebUI.

We have installed the ControlNet model for you by default, which can enhance your control over text-to-image and image-to-image generation through edge detection, depth detection, human pose recognition, etc., making the generated images more in line with your expectations.

We strongly recommend that customers run this solution on EC2 instances with GPUs, and we recommend that your EC2 instance type be no less than g4dn.2xlarge or g5.2xlarge. This AMI does not support ARM-based CPUs.