Stable Diffusion Glossary

Question 1

what is a Checkpoint model

Answer 1

A checkpoint model is a more precise name for a Stable Diffusion model. It is used to distinguish from LoRA, textual inversion and Lycoris.

Question 2

What is a CFG scale

Answer 2

The Classifier-Free Guidance scale controls how much the prompt should be followed in txt2img and img2img.

Question 3

What is 4x-Ultrasharp (2)

Answer 3

4x-ultrasharp is a popular AI upscaler that produces sharp images.
It is popular among Stable Diffusion users.

Question 4

What is an AI upscaler

Answer 4

An AI upscaler is an AI model that enlarges an image while adding details.

Question 5

What is an Ancestral sampler (3)

Answer 5

1. An ancestral sampler adds noise to the image at each sampling step.
2. They are stochastic samplers because the sampling outcome has some randomness to it.
3. They usually have a standalone letter “a” in their name. E.g. Euler a.

Question 6

What is AnimateDiff (2)

Answer 6

1. AnimateDiff is a text-to-video method for Stable Diffusion.
2. It uses a motion control model to influence a Stable Diffusion model to generate a video as a sequence of images with motions.

Question 7

What is Anything v3

Answer 7

Anything v3 is a celebrated anime-style Stable Diffusion model. It is a Stable Diffusion v1.5 model.

Question 8

What is AUTOMATIC1111(3)

Answer 8

1. AUTOMATIC1111 is a popular open-source, community-developed user interface for stable diffusion.
2. AUTOMATIC1111 is the name of the user who started the project.
3. The official project name is Stable Diffusion Web UI.

Question 9

What is Civitai, and how can it be used? (4)

Answer 9

1. Civitai is a website that holds a large number of Stable Diffusion models.
2. You can use the AUTOMATIC1111 extension Civitai Helper to facilitate the download.
3. Compared to Hugging Face, Civitai specializes in Stable Diffusion models.
4. You can see many user-generated images there.

Question 10

What is ComfyUI (2)

Answer 10

1. ComfyUI is a node-based user interface for Stable Diffusion.
2. It is popular among advanced Stable Diffusion users.

Question 11

ControlNet (5)

Answer 11

1. ControlNet is a neural network that controls image generation
2. **by adding extra conditions. **
3. You can use it to control human poses and image compositions.
4. It is a major breakthrough in Stable Diffusion.
5. There is :
   ControlNet for v1 models
   ControlNet for SDXL models

Question 12

What is DDIM

Answer 12

Denoising Diffusion Implicit Models (DDIM) is one of the first samplers for solving diffusion models.

Question 13

What is Denoising Strength

Answer 13

Denoising strength controls how much the image should change in the img2img process.

Question 14

Diffusion

Answer 14

1. Diffusion is an AI image-generation technique starting with a random image and gradually denoising it to a clear image.
2. It is inspired by the Langevin dynamics formulation of the diffusion process in Physics.

Question 15

WHat is DPM solver

Answer 15

Diffusion Probabilistic Model Solvers (DPM-Solvers) belong to a family of newly developed solvers for diffusion models.

Question 16

What is Dreambooth (3)

Answer 16

1. Dreambooth is a training technique to modify a checkpoint model.
2. Needing** as few as 5 images**, you can use it to inject a person or a style into a model.
3. A dreambooth model needs a trigger keyword in the prompt to trigger the injected subject or style.

Question 17

What is EMA

Answer 17

1. EMA stands for **Exponential Moving Average. **
2. In a Stable Diffusion model, it is **the average weights over the last training steps. **
3. Instead of the last training step,** a checkpoint model often use the EMA weights to improve stability**.

Question 18

Embedding

Answer 18

1. An embedding is** the product of textual inversion.**
2. It is a small file for modifying an image.
3. You apply embedding by putting in the associated keyword in the prompt or negative prompt.

Question 19

Extension (2)

Answer 19

1. An **extension extends the functionality of AUTOMATIC1111 WebUI. **
2. eg, ControlNet is implemented through an extension.

Question 20

What is Euler

Answer 20

The Euler method is the simplest sampling method for solving a diffusion model.

Question 21

What is Fooocus (3)

Answer 21

1. Fooocus is a Stable Diffusion software designed for simplicity.
2. It centers the user experience on prompting and image generation.
3. It’s free and open source.

Question 22

What is Hugging Face (3)

Answer 22

1. Hugging Face is a website that hosts a large amount of AI models.
2. In addition, they develop tools to help run and host the models.
3. Compared to Civitai, Hugging Face covers all AI models, not just Stable Diffusion.

Question 23

What is Hypernetwork (2)

Answer 23

1. A hypernetowrk is a small neural network that modifies the cross-attention module of the U-net noise predictor.
2. Similar to LoRAs and embeddings, they are small model files used for modifying a checkpoint model.

Question 24

What is the Karras noise schedule

Answer 24

1. Karras is a noise schedule proposed in the Karras article, which studied a unified framework for denoising images in diffusion AI models.

Question 25

What is K-diffusion/K-sampler

Answer 25

K-diffusion or K-samplers refer to sampling methods that Katherine Crowson’s k-diffusion GitHub repository implemented.

Question 26

What is Latent diffusion

Answer 26

The Latent diffusion refers to a diffusion process in the latent space. For example, Stable Diffusion is a latent diffusion model.

Question 27

What is LDM

Answer 27

The latent Diffusion Model (LDM) is an AI model that performs diffusion in the latent space.