AI-Generated Art
Art created using artificial intelligence models like Stable Diffusion, which generate images based on text prompts.
AnimateDiff
An extension for Stable Diffusion that enables the creation of animations and videos from text prompts or by modifying existing videos. Learn more: AnimateDiff
API (Application Programming Interface)
A set of functions allowing interaction with RunDiffusion services, including integration with Photoshop plugins.
API Mode
A setting in RunDiffusion that allows external applications and scripts to interact with the AI image generation service. Learn more: API and the Photoshop Plugin
Automatic1111
A popular web UI for running Stable Diffusion models with advanced customization options.
Batch Count
Generates images sequentially.
Batch Processing
Running multiple images at once using AI models, often with Img2Img or ControlNet. Learn more: Img2Img Batch Processing
Batch SizeBatch Size generates multiple images at once (uses more VRAM).
CFG (Classifier-Free Guidance) Scale
A parameter in Stable Diffusion that controls how closely the generated image adheres to the text prompt. A higher CFG scale makes the output more aligned with the prompt, while a lower scale allows for more creativity.
Checkpoint (Model Checkpoint)
A stored state of a trained model, used for generating images in Stable Diffusion. Learn more: Checkpoint Merging
Checkpoint Loader
A node that Loads the main model (like SDXL, AnythingV5, etc.).
CivitAI
A website where users can download custom models and LoRAs. Learn more: How to Download from CivitAI on RunDiffusion
Clip Skip
Adjusts how many layers of the CLIP model are skipped. Lower values improve accuracy, while higher values may reduce artifacts. (Common: 1 or 2)
CLIP Text Encode
Node that Converts text prompts into a format the AI understands.
ComfyUI
A node-based workflow interface for Stable Diffusion, allowing advanced customization. Learn more: ComfyUI Basics
Configuration File
Configuration files contain the settings for each application. These configuration files can be edited. Reset and Restart: A guide to deleting config files
ControlNet
An extension for Stable Diffusion that enables more precise control over generated images by using reference images, depth maps, or edge detection. Learn more: ControlNet
Corrupt Model
A damaged or incomplete AI model file that fails to load properly, often requiring reinstallation or redownloading.
CUDA Memory Error
A common GPU-related error in AI image generation caused by excessive memory usage, often requiring optimizations or reduced batch sizes. Learn more: Troubleshooting and Common Errors
Deforum
A framework for creating AI-generated video animations using Stable Diffusion. Learn more: Deforum: The Wild West of AI Video
Denoise Strength
Controls how much an image changes in img2img or upscaling.
Extensions
Plugins that add new features to applications, such as Deforum for animations or wildcard generators.
Face Restoration
A tool in Fooocus that enhances facial details in AI-generated images, reducing distortions and improving realism.
Face Swapping
A technique for replacing faces in AI-generated images using Stable Diffusion tools. Learn more: Face Swapping on RunDiffusion
Fooocus
A streamlined AI image generation interface designed for simplicity and high-quality outputs. It automates complex settings like prompt optimization and image refinement, making it easier for users to create AI-generated art with minimal adjustments. Learn more: Fooocus Guide
Filebrowser
A built-in tool in RunDiffusion for managing stored files, including models, LoRAs, and generated images. Learn more: Navigating the File Explorer
Guidance Scale
Controls how strictly the AI follows your text prompt in Stable Diffusion.
Hires Fix
(High-Resolution Fix) Upscales images for better detail without distortion.
Hypernetwork
A small neural network that modifies AI model behavior without changing the base model itself. Learn more: LoRA, Hypernetworks, and Embeddings in Automatic1111
Img2Img
A feature in Stable Diffusion that allows modifying an existing image using AI-generated elements. Learn more: Img2Img
Inpainting
A technique used to edit or fill missing parts of an image using AI. Learn more: Runnits
Instant ID
A tool that allows the creation of AI-generated portraits with a recognizable identity. Learn more: Create beautiful stylized portraits with Instant ID!
JSON File
File format used to save and share ComfyUI workflows
Kohya
A set of tools used for LoRA Training, helping fine-tune AI models efficiently. Learn more: Basic LoRA Training with Kohya
KSampler
The main node responsible for generating an image. Typically controls Sampling method, Steps and CFG Scale.
Latent Image
An intermediate image representation before the final render. It exists in a compressed form used by the AI.
LoRA (Low-Rank Adaptation)
A technique that allows fine-tuning of AI models with minimal data. Learn more: Flux LoRA Training
Model Merging
Combining multiple AI models to create unique outputs. Learn more: Model Merging
Mismatched Models
A compatibility issue that arises when using AI models that are not designed to work together, leading to unexpected results.
Negative Prompting
A technique used to guide AI-generated images by specifying elements to avoid. Learn More: Negative Prompting
Nodes
The building blocks of ComfyUI. Each node performs a specific function, like loading a model, generating an image, or applying effects.
Outpainting
Extending the edges of an image using AI-generated content. Learn more: Outpainting with Fooocus
PBRemTools
A set of tools in Automatic1111 that enhance AI-generated image refinement. Learn More: Unlock the power of PBRemTools in Automatic111
Prompt
A text-based instruction given to an AI model to generate an image. A well-crafted prompt includes descriptive details, styles, and modifiers to achieve the desired output. Learn More: Prompt Guide
Prompting
The process of writing detailed text descriptions to generate images in AI models. Learn More: Prompt Guide
Quantization
Reducing model size to improve efficiency
Refiner Model
A secondary AI model used to enhance fine details, such as textures, faces, and backgrounds, after the initial image generation.
Restart UI
A troubleshooting step used to refresh the RunDiffusion interface and resolve common errors.
Resolution
The size of the generated image, measured in pixels. Higher resolutions result in more detailed images but require more processing power.
RunDiffusion
A cloud-based platform that provides access to Stable Diffusion, Automatic1111, and other AI tools. Learn: Getting Started with RunDiffusion
RunDiffusion Learn
A collection of tutorials and guides for RunDiffusion Services, Applications and Runnits. Learn More: RunDiffusion Learn
Runnit Platform
A video-generation tool available on RunDiffusion, designed for creating AI-based animations. Learn More: Runnits
Runnit Board
A platform that aggregates various open-source tools into a single workspace, allowing users to combine and utilize them within a unified environment. Runnit Board enables users to stack multiple tools, create workflows connecting them, and facilitate ideation and testing across different tools to achieve creative goals. It streamlines AI-assisted workflows by integrating multiple functionalities in one place.
Runnit Tool
Within the Runnit Platform, a Runnit Tool (or simply Tool) refers to a single-purpose utility designed to help users create or accomplish a specific, desired outcome. For example, Background Remover is a Runnit Tool.
Sampling Methods
Different techniques used to generate images in AI models, impacting speed and quality. Learn More: Sampling Methods
Seed
A number that determines how an image is generated. Keep the same seed for consistency, change for variations
Stable Diffusion
An open-source AI model for generating images from text prompts, widely used in platforms like RunDiffusion.
Steps
Number of times the AI refines an image. The more steps the better quality but slower generation.
StorageAllocated space in RunDiffusion for saving custom models, LoRAs, and generated images
Style Presets
Predefined artistic styles in Fooocus that adjust the AI’s rendering to create specific looks, such as realistic, anime, cinematic, or painterly.
TXT2IMG
A feature in Stable Diffusion that generates an image purely from a text description. Learn more: TXT2IMG
Timeout Error
An issue that occurs when the system takes too long to respond, often due to high computational demand or server load.
Team Plan
A subscription option for collaborative AI image generation, offering shared resources and enhanced storage. Learn More: How to Use Your Team Plan
Tensor
A multi-dimensional array used to process image data within the AI model.
Upscaling
Enhancing the resolution of an AI-generated image while maintaining detail and quality. Learn More: Upscaling 101
VAE (Variational Autoencoder)
Converts latent images into actual images with color and detail.
Wildcards
A method of using variables in text prompts to generate a variety of outputs with randomized elements. Learn More: Wildcards!
Workflows
A collection of connected nodes that define an image generation process. You can create and modify workflows for different styles and effects.
XYZ Plot
A Script for comparing multiple settings in AI image generation frequently used to test checkpoints and lora.
Y-Parameter (XYZ Plot)
Used in XYZ Grids to compare different AI settings like CFG, steps, or samplers.
Z-Scale
Zoom Factor in Deforum. Used in Deforum animations to control zoom effects.
