718
Image generators like Stable Diffusion can support you in your creative work. Artificial intelligences are enjoying great popularity, also when it comes to the topic of image generation.
Stable Diffusion: Overview
Stable Diffusion uses open source software that allows you to build on and work with existing code. The application is based on the Deep Learning system, which is a sub-discipline of Machine Learning and deals with the development of neural networks.
- Stable Diffusion: This is an image generator. With the help of an AI program, Stable Diffusion creates images from your sentences. Stable Diffusion 1 was released in August 2022, and its successor, Stable Diffusion 2, was released the following December.
- Deep learning system: Conventional neural networks consist of one or a few layers. In contrast, deep neural networks consist of multiple layers, which allows them to recognise and learn complex patterns and relationships in large amounts of data. Deep learning is often used in applications such as image recognition or speech recognition.
- Company: Stable Diffusion was developed by Stability AI. This is a London-based start-up that was founded in 2020. With CEO Emad Mostaque, the start-up managed to generate millions of dollars in investments within a short period of time.
- Contributors: Several parties participated in Stable Diffusion. Among them are Stability AI as well as RunwayML, EleutherAI, LAION and a research group from LMU Munich.
- Advantages of Stable Diffusion: Compared to competitors like DALL-E or Midjourney, Stability AI’s tool is free of charge. In addition, Stable Diffusion has a higher resolution in direct comparison with DALL-E.
Stable diffusion: prompts to the tool
To use the image generator, you must enter so-called “prompts”. These are prompts that you give to the AI. The AI thus converts your texts into images. You can also write “negative prompts” to tell the AI what not to integrate images. To get the best possible results, proceed as follows:
- Be as detailed as possible when describing the subject.
- Use keywords rather than worded box sentences. This is easier for the AI to process.
- Also feel free to include a description of the background/environment or colour preferences.
- Describe the style you would like to give your image, such as an oil painting, minimalist, retro, 3D illustration, and more. You can use the negative prompts to exclude certain styles.
- For example, specify the mood you want the image to convey.
