Krita AI diffusion, A look at v1.15 update, bugs on older hardware and new cool features

Streamtabulous
9 Mar 202428:18

TLDRThe video discusses the AI diffusion 1.15 update by akley for Krita, highlighting new features that enhance artwork editing capabilities, rivaling Adobe Photoshop's generative fill. It covers the update process, bug issues on older hardware, and potential fixes. The creator also mentions the new lightning model by Civ AI, emphasizing the model's versatility and the community's shared prompts for testing.

Takeaways

  • ๐ŸŽจ The AI diffusion 1.15 update by akley for Cryer brings new features that enhance the editing capabilities of artwork, making it more powerful and on par with Adobe Photoshop's generative fill.
  • ๐Ÿ” Users can now have more control over the generative process, which helps prevent glitching and allows for better customization of the artwork.
  • ๐Ÿ“š The update includes the ability to copy prompts and a suggestion for future versions to include the ability to save prompts as text files and import them directly.
  • ๐Ÿ”ง A new feature allows users to discard rendered parts they don't like by right-clicking, providing more flexibility during the editing process.
  • ๐Ÿ”„ The update introduces a 'generate custom' option, which considers the entire image for context, leading to more accurate and controlled image generation.
  • ๐Ÿ–Œ There's a mention of a new 'lightning model' by Civ AI that works well for both fantasy creatures and humans, showcasing the versatility of the tool.
  • ๐Ÿ’พ The process for updating the AI diffusion is straightforward, involving downloading a zip file and extracting it to overwrite the existing files.
  • ๐Ÿ–ฅ๏ธ The video discusses bugs encountered when using the AI diffusion on older hardware like the GTX 1070, suggesting that there might be memory management issues.
  • ๐Ÿ› ๏ธ A workaround for the generative expand error involves adding a dot to the image, which tricks the AI into processing the entire image correctly.
  • ๐Ÿ“‰ The speaker expresses a desire for lower RAM settings or a checkbox to input arguments that could fix the bugs, indicating a need for optimization for different hardware capabilities.
  • โš™๏ธ Despite the bugs, the update is praised for its improvements and the potential it holds for photo restoration and digital artwork, even on older hardware.
  • ๐Ÿ“ˆ The video ends with an encouragement for viewers to like, subscribe, and share the content for better visibility and to help others facing similar issues.

Q & A

  • What is the main focus of the Krita AI diffusion 1.15 update by akley for Cryer?

    -The Krita AI diffusion 1.15 update focuses on introducing new features that make the software more powerful and versatile, allowing users to edit their artwork in ways that were not possible before. The update brings it on par with Adobe's Photoshop for generative fill, with the added benefit of more control to prevent glitching.

  • How can users update their Krita AI diffusion to version 1.15?

    -To update Krita AI diffusion to version 1.15, users can visit the GitHub page of akley, download the zip file, extract it, and overwrite the existing files. Upon opening Krita, the software will prompt the user to update the configuration for the AI, ensuring all new features are included.

  • What is the significance of the lightning model introduced by Civ AI?

    -The lightning model introduced by Civ AI is designed for generating fantasy creatures but has proven effective for rendering humans as well. It offers a new option for users to experiment with and has been well-received by the community, with some users sharing their prompt information for others to try.

  • What new core components and features does the AI diffusion 1.15 version bring to Krita?

    -The AI diffusion 1.15 version introduces several new core components and features, such as the ability to discard rendered elements, copy prompts, and an enhanced auto-selection tool. It also includes new samplers like DPM Plus+, SD, and sgm, which can improve the quality and speed of rendering.

  • How does the 'expand' feature in the 1.15 update work?

    -The 'expand' feature allows users to extend an image in a way that takes the entire image into account, rather than just the prompt information. This results in a more controlled and coherent expansion, avoiding random generation and ensuring a better blend with the original image.

  • What is the issue encountered when trying to use the face tool in the AI diffusion 1.15 update?

    -Users may encounter a runtime error when using the face tool in the AI diffusion 1.15 update, particularly on older hardware. This error is related to a mismatch between expected data types, which may be resolved by adding specific arguments in the configuration file or by upgrading the hardware.

  • What is a workaround for the expand feature bug in the AI diffusion 1.15 update?

    -A workaround for the expand feature bug involves placing a dot within the area to be expanded. This provides the AI with enough information to generate the expansion while avoiding errors related to insufficient data for the generative process.

  • What are the implications of using older hardware like the GTX 1070 with the AI diffusion 1.15 update?

    -Using older hardware like the GTX 1070 with the AI diffusion 1.15 update may result in more bugs and slower performance. However, it is still possible to use the software effectively, and there may be workarounds for some issues. Users may need to adjust settings or consider upgrading their hardware for a better experience.

  • How does the AI diffusion 1.15 update enhance photo restorations?

    -The AI diffusion 1.15 update enhances photo restorations by providing more control over the generative process, allowing users to remove or add content more easily. The update's features, such as auto-selection and expansion, make it more efficient and effective for photo restoration tasks.

  • What feedback does the speaker provide for the developers of the AI diffusion plugin?

    -The speaker suggests that the developers could add a feature to include arguments in the settings to fix certain bugs and improve usability. They also express hope that the developers will address the issues faced on older hardware, allowing users to continue using the software effectively.

Outlines

00:00

๐ŸŽฅ Introduction to AI Diffusion 1.15 Update

The video begins with a warm welcome and an introduction to the AI Diffusion 1.15 update by akley for Cryer. The host highlights the new features of the update, emphasizing its enhanced capabilities that now rival Adobe's Photoshop in terms of generative fill, but with more user control to prevent glitching. The video also mentions a link to a tutorial on how to update the software and gives a shoutout to Civ AI for their newly released furnace model, which is capable of rendering both fantasy creatures and humans effectively.

05:01

๐Ÿ› ๏ธ Exploring the New Features and Workflows

The host delves into the new features of the AI Diffusion 1.15 update, discussing the ability to auto-select with outlining and the inclusion of various samplers like DPM Plus+, SD, and SGM. The video touches on the desire to see more samplers in future updates and the potential reasons behind their absence. The host also explains the use of the lightning sampler, which is designed to work seamlessly with the lightning models for better performance and ease of use.

10:02

๐ŸŽจ Demonstrating the Customization and Editing Capabilities

The video provides a practical demonstration of the AI Diffusion's customization and editing capabilities. The host shows how to select and modify specific parts of an image, such as discarding unwanted elements and copying prompts. The focus is on the new options for content generation, selection borders, and masks, as well as the ability to adjust the AI's behavior with settings like 'neutral' and 'imp paint' for seamless blending and improved rendering results.

15:03

๐Ÿž Discussing Bugs and Their Workarounds

The host discusses encountering bugs in the AI Diffusion 1.15 update, sharing specific error messages and their potential causes. The video explores workarounds for these issues, such as adding a 'magic dot' to resolve generative expand errors. The host expresses hope that the plugin's creator, suspicious Firefly, will address these bugs in future updates. The video also acknowledges the limitations of older hardware and the host's personal challenges with upgrading their system.

20:05

๐Ÿ“ข Final Thoughts and Call to Action

In the concluding segment, the host shares final thoughts on the AI Diffusion 1.15 update, praising its powerful features and improved capabilities for photo restorations and digital art. The host encourages viewers to like, subscribe, and turn on notifications for future content, and invites them to share their experiences with the software, including hardware specifications, render times, and resolutions. The video ends with a thank you note for watching and a teaser for the next video.

Mindmap

Keywords

๐Ÿ’กAI diffusion

AI diffusion refers to the process of using artificial intelligence to generate and alter images or artwork. In the context of the video, it is the main focus of the 1.15 update by akley for Cryer, which introduces new features to enhance the user's ability to edit and manipulate artwork in ways that were not previously possible. The AI diffusion is compared to Adobe's Photoshop for its generative fill capabilities but with added control to prevent unwanted glitching or artifacts in the generated images.

๐Ÿ’กCryer tools

Cryer tools are a set of plugins or extensions that are used in conjunction with the AI diffusion update. These tools are designed to enhance the functionality and usability of the AI diffusion, providing features such as auto-selection and outlining. The video mentions that these tools are very handy and can be found on akley's GitHub page, indicating that they are likely open-source or freely available to users who wish to utilize them in their artwork creation process.

๐Ÿ’กGitHub

GitHub is a web-based hosting service for version control and collaboration that is used by developers to store and manage their code. In the video, the presenter mentions visiting akley's GitHub page to find the AI diffusion and Cryer tools, suggesting that the tools are likely stored there in a repository for users to download and install. This implies that the tools are likely open-source and community-driven, with updates and improvements being made by a collective of contributors.

๐Ÿ’กgenerative fill

Generative fill is a feature in image editing software that uses algorithms to fill in selected areas with new content. In the context of the video, the presenter compares the AI diffusion's generative fill to that of Adobe Photoshop, highlighting that the AI diffusion offers a similar capability but with more user control. This feature is particularly useful for tasks such as photo restoration or digital painting, where users may want to seamlessly integrate new elements into an existing image.

๐Ÿ’กglitching

Glitching in the context of digital art refers to unintended visual artifacts or errors that occur in the image generation process. The video discusses how the new features in the AI diffusion 1.15 update help to prevent certain types of glitching, thereby improving the quality and reliability of the generated images. This is an important aspect of the update, as it addresses a common issue that users may have experienced with previous versions of the AI diffusion tool.

๐Ÿ’กauto-selection

Auto-selection is a feature that automatically identifies and selects specific areas of an image based on certain criteria or characteristics. In the video, it is mentioned as a handy feature of the Cryer tools that work in conjunction with the AI diffusion. This functionality can greatly streamline the editing process by reducing the need for manual selection and allowing users to focus on the creative aspects of their work.

๐Ÿ’กfurnace Model

The furnace Model, as mentioned in the video, is a new model released by Civ AI that is designed for generating fantasy creatures but can also be used for rendering humans. The presenter expresses excitement about this model and its potential applications, indicating that it is a significant addition to the AI diffusion tool's capabilities. The model's versatility and the positive feedback from users who have shared their renders demonstrate its value in enhancing the creative possibilities for artists using the AI diffusion tool.

๐Ÿ’กprompt information

Prompt information refers to the input or instructions provided to the AI diffusion tool to guide the generation of images. In the video, the presenter discusses the ability to copy and reuse prompts, which is a convenient feature for artists who want to experiment with different variations of a scene or subject. The prompt information is crucial for achieving the desired outcome in the generated images, as it informs the AI about the specific elements or characteristics that should be included.

๐Ÿ’กsampler

A sampler in the context of AI diffusion is a method or algorithm used to generate the content for the images. The video mentions different types of samplers, such as DPM Plus+, SD, and sgm sampler, which are used to refine the generation process and achieve different visual effects. The presenter expresses a desire to see more samplers implemented in future updates, as they can significantly impact the quality and appearance of the rendered images.

๐Ÿ’กrender time

Render time refers to the duration it takes for the AI diffusion tool to process and generate an image. The video discusses the render times on older hardware, such as the GTX 1070, and how it compares to the performance of the tool on newer or more powerful hardware. The presenter notes that while the tool is powerful and offers many new features, the render times can be longer on older hardware, which may limit the ability of some users to fully utilize the tool's capabilities.

๐Ÿ’กbugs

Bugs are errors or issues that occur in software that can cause unexpected behavior or prevent the software from functioning as intended. In the video, the presenter encounters bugs when using the AI diffusion 1.15 update on older hardware, such as the GTX 1070. These bugs manifest as runtime errors and can prevent certain features from working correctly. The presenter hopes that the developer, suspicious Firefly, can address these issues in future updates to improve the stability and performance of the tool on a wider range of hardware.

Highlights

Krita AI diffusion v1.15 update introduces powerful new features, making it on par with Adobe Photoshop's generative fill.

The update allows for better control to prevent glitching in artwork edits.

Auto-selection with outlining feature is now available for more precise editing.

AI diffusion can now be updated easily by downloading a zip file and overwriting the existing files.

Civ AI has released a lightning model designed for rendering fantasy creatures, but also works well for humans.

The new version includes the ability to copy and paste prompts for easier editing and generation.

The lightning sampler is introduced, which is optimized for use with the new models.

The DPM Plus+, SD, and SGM samplers are not yet available in the new version, but are highly anticipated.

The auto-selection feature expands images by considering the entire image, leading to more accurate fills.

The generate custom function allows for specific content selection and border masking for more control in image generation.

The update includes options to adjust the diffusion and paint outside of selections for better blending.

The ability to remove content from an image has been improved, making photo restoration easier.

The update has introduced some bugs on older hardware, such as the GTX 1070.

A workaround for the expand error involves adding a dot to the image before processing.

Despite the bugs, the new features in v1.15 are highly appreciated and offer significant improvements.

The video creator plans to reach out to the plugin developer, supicious Firefly, to address the bugs found.

The new update allows for photo restoration and digital painting with more control and ease than before.

The video encourages viewers to share their experiences with the new update and their hardware configurations.