AUTOMATIC1111のおすすめ拡張機能9選んんんほおお!【Stable Diffusion】

テルルとロビン【てるろび】旧やすらぼ
30 Jun 202310:55

TLDRRobinとTeruruが紹介するAutomatic1111の9つの便利な拡張機能について語る。Image Browserで出力フォルダーの画像をブラウジングし、メタデータ検索やランキング機能で整理。タグ補完のTag Complete、システム情報をリアルタイム表示するSystem Info、テンプレート編集のStyle Editor、アスペクト比の計算を助けるAspectRatio SelectorとHelper、設定を保存するConfig Preset、顔の詳細を追加するAfter Detailer、セグメント編集のInPaint Anythingなどが挙げられる。これらの機能は、ユーザーが生成と設定の煩わしさを軽減し、操作を効率化できると述べている。

Takeaways

  • 🖼️ The Image Browser extension allows users to easily view and manage images in the Outputs Folder through a web UI, including searching and sorting by metadata.
  • 🔍 Tag Complete is a prompt auto-complete feature that suggests Danbooru-tags and helps with inputting prompts more efficiently.
  • 💻 System Info provides real-time information about the machine's current state, including VRAM usage and a list of recognized models and learning files.
  • 📝 Style Editor enables easy editing and management of saved prompt templates directly from the web UI, streamlining the process of organizing styles.
  • 🎨 Aspect Ratio Selector simplifies the process of generating images at specific sizes while maintaining their aspect ratio, with preset buttons and a calculator for custom ratios.
  • 📐 Aspect Ratio Helper is a slide bar auxiliary function that helps maintain the aspect ratio when adjusting the resolution, offering a different approach from the selector.
  • 📑 Config Preset saves frequently used prompt and sampling settings, allowing for quick recall and reducing setup time.
  • 🔎 After Detailer is an extension that automatically enhances the details of an illustration's face, providing a more refined output with minimal effort.
  • 🖌️ InPaint Anything uses the Segment Anything model to modify illustrations and create masks for specific parts of an image, offering various processing options.
  • 📈 The ranking function in Image Browser lets users rank images and then sort the view to display only top-ranked images, which is useful for sorting through mass-produced images.
  • 📚 The Style.csv file, where saved styles are stored, can now be easily edited with the Style Editor, eliminating the need to manually edit the file in a text editor.

Q & A

  • What is the first expansion function introduced for Automatic1111?

    -The first expansion function introduced is the Image Browser, which allows users to easily browse the output images on the web UI without having to open the Outputs Folder and check each image individually.

  • How does the Image Browser function simplify the process of viewing and managing output images?

    -The Image Browser automatically reads and displays images in the Outputs Folder, enabling users to view, search metadata, delete, and send images to the control net or trash can directly from the web UI.

  • What is the unique feature of the Image Browser that assists in sorting images?

    -The unique feature of the Image Browser is the ranking function, which allows users to rank images they consider good and then sort by rank to display only those images.

  • How does the Tag Complete function assist users in inputting prompts?

    -Tag Complete is a prompt auto-complete function that suggests Danbooru-tags, corrects typing errors, and provides related word suggestions, making prompt input easier and more efficient.

  • What is the System Info function and what does it display?

    -System Info is a function that displays the current state of the machine in real time, including the consumption of VRAM during generation and a list of recognized models and learning files.

  • How does the Style Editor make managing saved prompts easier?

    -The Style Editor allows users to edit the Style.csv file on the web UI, making it simple to view, edit, or delete registered styles without needing to open the file in a notepad.

  • What is the purpose of the Aspect Ratio Selector and how does it help with image generation?

    -The Aspect Ratio Selector helps users generate images of a certain size while maintaining the desired aspect ratio. It includes preset buttons for common ratios and a calculator to assist with ratio adjustments.

  • How does the Aspect Ratio Helper differ from the Aspect Ratio Selector?

    -The Aspect Ratio Helper is an auxiliary function of the slide bar that adds a small command next to the resolution setting with preset aspect ratios, allowing users to adjust the bar while maintaining the ratio.

  • What is the Config Preset and how does it benefit users?

    -The Config Preset allows users to save their frequently used prompt and sampling settings, making it easy to quickly recall these settings when starting up, which can significantly reduce the hassle of generating and setting up.

  • What is the After Detailer extension and how does it enhance illustrations?

    -After Detailer is an extension that automatically detects the face in an illustration and adds more details to it. It can be used with T2i and I2i and provides a noticeable enhancement in detail with minimal operation.

  • How does the InPaint Anything feature work and what are its main uses?

    -InPaint Anything is a feature that uses an object detection model to modify illustrations and create masks for specific parts. It allows users to select a model, segment an image, create a mask for a specific object, and then process the image using options like Control Net or Mask Only for various editing purposes.

  • What are the three types of models available in the InPaint Anything feature?

    -The three types of models available in the InPaint Anything feature are the original Facebook Segment Anything model, an improved high-quality model, and a high-speed model.

Outlines

00:00

🖼️ Image Browser Overview

The first paragraph introduces the Image Browser, an expansion function of Automatic1111 that simplifies the process of browsing output images. Instead of manually checking each image in the Outputs Folder, this tool automatically displays images on the web UI. Users can easily browse through the images, view metadata, and perform actions such as deleting or sending images to the control net. A standout feature is the search functionality within the metadata, which allows users to filter images based on specific prompts, model names, sampling methods, and expansion function names. Additionally, the ranking function enables users to sort images based on their assigned rank, which is particularly useful for organizing images after mass production with Generate Forever. The installation process is straightforward, involving the familiar extension method.

05:01

📝 Tag Complete and System Info

The second paragraph covers two distinct features: Tag Complete and System Info. Tag Complete is an auto-complete tool for prompts, specifically designed to suggest Danbooru-tags used on image posting sites like Danbooru. It assists with typing errors and related word suggestions, enhancing the ease of prompt input. The tool also displays the number of hits for each tag, indicating the popularity and potential relevance of the content. Users are advised to increase the default number of suggestions for a better experience. System Info, on the other hand, is aimed at users interested in the system's current state. It provides real-time information about the machine, including VRAM consumption during image generation. Furthermore, it lists recognized models and learning files, simplifying the process of locating these resources without the need to manually check folders.

10:03

🎨 Style Editor and Aspect Ratio Tools

The third paragraph discusses the Style Editor and various Aspect Ratio tools. The Style Editor is an extension that allows users to manage their saved prompt templates more efficiently. Instead of manually editing a Style.csv file, users can now edit and delete registered styles directly through the web UI. This feature simplifies the process of managing saved styles and is particularly useful for users with many saved templates. The Aspect Ratio Selector and Aspect Ratio Helper are tools designed to assist with generating images of specific sizes while maintaining their aspect ratios. The Selector provides preset buttons for common ratios and a simple calculator to determine the necessary changes in aspect ratio when altering resolutions. The Helper, conversely, operates as a slide bar auxiliary, allowing users to adjust the resolution while keeping the aspect ratio constant. Both tools are aimed at making the process of setting aspect ratios more intuitive and less error-prone.

🖊️ Config Preset and After Detailer

The fourth paragraph highlights the Config Preset and After Detailer. Config Preset is a feature that saves users the trouble of re-entering frequently used prompt and sampling settings by allowing them to save and quickly recall these configurations. It adds a pull-down menu for easy access to presets, which can be customized and managed through an Add Remove button. This tool is particularly useful for users who often reuse certain settings, streamlining the generation process. After Detailer is an extension feature that automatically enhances the details of an illustration's face. It works with T2i and I2i and can significantly improve the quality of facial details with minimal effort. Users can also create variations in facial expressions by entering specific prompts for After Detailer, offering a powerful yet easy-to-use enhancement to the default settings.

🖼️ InPaint Anything and Mask Only

The fifth and final paragraph introduces InPaint Anything, an extension based on Facebook's object detection model, Segment Anything. This feature enables users to modify illustrations and create masks for specific segments. It offers a range of models to choose from, including a high-quality model and a high-speed model, catering to different machine capabilities. The process involves selecting a model, downloading it, and then segmenting the image. Users can create masks for specific objects and apply various processing methods, such as InPaint, Cleaner, Control Net, and Mask Only. The Control Net mode allows for the use of a preferred model to modify the selected segment, while Mask Only extracts just the segment mask, useful for creating collage materials or extracting parts. This tool provides a natural finish similar to the standard i2i InPaint and offers a simple yet versatile method for image editing.

Mindmap

Keywords

💡Automatic1111

Automatic1111 is a software or tool that offers various expansion functions to enhance its capabilities. In the video, it is the central theme as the host introduces different useful functions that can be added to this tool to improve the user experience. It is mentioned as the main subject of the video's discussion.

💡Image Browser

The Image Browser is an expansion function that allows users to easily browse and manage images in the Outputs Folder through a web UI. It simplifies the process of checking output images by automatically reading and displaying them. It is highlighted in the video as a simple yet powerful tool for users to interact with their generated images.

💡Metadata

Metadata refers to the data that provides information about other data. In the context of the video, it is used to describe additional information about the images such as the prompt used, model name, sampling method, and the name of the expansion function. The Image Browser's ability to search and sort by metadata is emphasized as a key feature.

💡Tag Complete

Tag Complete is an auto-complete feature for prompts, specifically designed to suggest Danbooru-tags, which are used on the image posting site Danbooru. It aids in inputting prompts by offering suggestions for typing errors and related words, thus facilitating the creation process. The video demonstrates how this feature can increase the efficiency of prompt input.

💡System Info

System Info is a function that displays real-time information about the machine's current state. It is useful for users who want to monitor VRAM consumption during image generation. Additionally, it lists recognized models and learning files, providing a convenient overview of the system's status without the need to manually check folders.

💡Style Editor

The Style Editor is an extension that enables users to edit the Style.csv file, which contains saved prompt templates, directly through the web UI. This feature simplifies the management of saved styles by allowing users to easily add, edit, or delete styles, making it a convenient tool for streamlining the template management process.

💡Aspect Ratio Selector

The Aspect Ratio Selector is a tool that assists users in generating images of a specific size while maintaining the desired aspect ratio. It simplifies calculations by providing preset buttons and a calculator to determine the necessary adjustments for width and height, ensuring that the output matches the user's requirements.

💡Aspect Ratio Helper

The Aspect Ratio Helper is a slide bar auxiliary function that complements the Aspect Ratio Selector. It allows users to adjust the resolution while keeping the aspect ratio constant through a preset command next to the resolution setting. It offers a different approach to maintaining aspect ratio, providing users with options to suit their preferences.

💡Config Preset

Config Preset is an extension that allows users to save and quickly recall their frequently used prompt and sampling settings. It adds a pull-down menu for easy access to presets, streamlining the process of starting up the tool with preferred settings, and reducing the hassle of reconfiguring settings each time.

💡After Detailer

After Detailer is an extension feature known for automatically detecting the face in an illustration and adding more details to it. It enhances the quality of generated images by refining facial features, providing a more detailed and polished output. The video demonstrates its effectiveness with an example of an illustration where facial details are improved.

💡InPaint Anything

InPaint Anything is a feature that utilizes an object detection model to modify illustrations and create masks for specific parts of an image. It operates through a simple flowchart-like interface, allowing users to select models, segment images, and process selected areas with various options like InPaint, Cleaner, Control Net, and Mask Only. The video showcases how it can be used to change the color of an object in an illustration, such as a tie, while maintaining a natural finish.

Highlights

Automatic1111 offers a variety of useful expansion functions for image and data management.

Image Browser allows for easy web UI browsing of output images and their metadata.

The Image Browser includes a search function for metadata, making it easy to find specific images.

Tag Complete is a prompt auto-complete feature that suggests tags based on popularity and relevance.

System Info provides real-time information about the machine's current state, including VRAM usage.

Style Editor enables easy editing and deletion of saved styles directly from the web UI.

Aspect Ratio Selector assists in maintaining image aspect ratios during generation.

Aspect Ratio Helper is a slide bar auxiliary function that maintains aspect ratio during adjustments.

Config Preset allows users to save and quickly recall prompt and sampling settings.

After Detailer is an extension that automatically enhances the details of faces in illustrations.

InPaint Anything uses an object detection model to modify illustrations and create masks.

The Image Browser can sort images by rank, aiding in organizing mass-produced images.

Tag Complete can increase prompt response and the likelihood of obtaining desired results.

System Info consolidates various system information, making it more accessible.

Style Editor simplifies the management of saved styles, enhancing user experience.

Aspect Ratio Selector includes a calculator for easy aspect ratio adjustments.

Config Preset offers a convenient way to save and access custom settings.

After Detailer can be used with T2i and I2i, providing a significant enhancement to image details.

InPaint Anything allows for easy extraction and modification of specific segments of an image.

The Image Browser is considered an essential expansion function for Automatic1111.