Stable-Diffusion
Stable Diffusion is a deep learning text-to-image diffusion model capable of generating photo-realistic images given any text input. In seconds you can create stunning artwork. Stable Diffusion uses a kind of diffusion model, called a latent diffusion model.
There are a variety of different models available which are installed using the Stable Diffusion Model Management (accessed by clicking the Model button in the Stable Diffusion dialog box).
Note the models are large downloads e.g. the Stable Diffsion 1.5 LCM model is 9.08 GiB, the Stable Diffusion X Base 1.0 is 13.24 GiB.
Here’s example output generated with the Stable Diffusion XL base 1.0 [Square][FP16] using the best performance power mode. I entered a simple text prompt: landscape picture with castle and moat
This image was literally the first image generated with that text prompt with 20 inference steps. On the Arc 140T iGPU a 1024×1024 image takes about 30 seconds to be generated.
Generation times are much quicker with other models. For example, with the Stable Diffusion 1.5 model it only takes 8 seconds to generate a 512×512 image. That’s only a few seconds longer than a mid-range dedicated graphics card such as the NVIDIA GeForce RTX 3060 Ti.
There’s support for negative prompts. They are the opposites of a prompt; they allow the user to tell the model what not to generate. Negative prompts often eliminate unwanted details like mangled hands or too many fingers or out of focus and blurry images.
Besides prompt to image, I can also perform image to image – this lets you upload a starting image or drawing which acts as the seed and generate a new image preserving the original’s shape and layout.
There’s also the option of inpainting which modifies a specific, masked area of an existing image by generating new content based on a text prompt and the image’s surrounding context.
The plugin doesn’t offer the versatility of software like Easy Diffusion, InvokeAI, Lama Cleaner, and others, but it still generates fast and impressive results with the integrated GPU.
Next page: Page 3 – Super Resolution, Semantic Segmentation
Pages in this article:
Page 1 – Introduction
Page 2 – Stable Diffusion
Page 3 – Super Resolution, Semantic Segmentation
Complete list of articles in this series:
| ASRock Industrial NUC BOX-255H | |
|---|---|
| Introduction | Introduction to the series and interrogation of the NUC BOX-255H |
| Benchmarks | Benchmarking the NUC BOX-255H |
| Power | Testing and comparing the power consumption under various workloads |
| Stable Diffusion | Deep Learning with Stable Diffusion |
| Audacity AI Plugins | Let's explore OpenVINO AI Plugins for Audacity |
| 3 Types of Cores | P-cores, E-cores and low power E-cores performance |
| GIMP AI Plugins | Stable Diffusion, Super Resolution, Semantic Segmentation |