Invoke - 3.6 Release - New User Interface Updates, and more.
TLDRInvoke AI has rebranded to Invoke, expanding its open-source software to include an Enterprise product for larger creative teams in gaming, media, and entertainment. The new user interface offers an improved experience with direct AI integration for professionals. The video provides a walkthrough of the updated features, including the options panel, work area, gallery, and advanced settings, highlighting the tools for text-to-image generation, image manipulation, and workflow optimization. The introduction of the workflow library and tiled upscaling workflow is also discussed, showcasing Invoke's commitment to enhancing creative AI solutions for professional use.
Takeaways
- 🎉 The company formerly known as Invoke AI has rebranded to Invoke, reflecting a focus on open-source software for creatives and the release of an Enterprise product for larger creative teams.
- 🚀 The Enterprise product is designed to help game studios, agencies, and media & entertainment companies securely use and deploy tools at scale across their teams.
- 🌟 Invoke's support from the Enterprise product allows for the continued development of great software for everyone and the release of their studio as a free open-source product.
- 💡 The user interface has been updated and is expected to evolve over time, with new AI features being integrated directly into the studio.
- 📌 The options panel provides powerful control capabilities, allowing users to switch between modalities, update generation settings, and access image actions.
- 🎨 The Gallery is divided into boards and images/assets, with boards capable of storing images and assets for use in various tools like control net or IP adapter.
- 📝 The text-to-image generation tab offers full control over prompts, with no hidden augmentation of terms, and features like dynamic prompts and embeddings for more precise control.
- 🔍 The image accordion allows control over image size and starting noise, with optimizations based on the selected model architecture.
- 🔧 Advanced options are available for less commonly used settings on a per-model basis, including features like seamless tiling and refiner controls.
- 🔄 The image-to-image tab has been largely unchanged, but now includes a use size button for matching the initial image's dimensions and ensuring coherent output.
- 🛠️ The workflow editor and workflow library have been introduced, allowing users to save and load workflows for streamlined content generation with predefined options.
Q & A
What is the new name of the company previously known as 'invoke AI'?
-The new name of the company is 'invoke'.
What is the purpose of the Enterprise product released by invoke?
-The Enterprise product is designed to help larger creative teams at game studios, agencies, and media and entertainment companies to use the tools securely and deploy them at scale across their teams.
How does the support for the Enterprise product contribute to invoke's development efforts?
-The support from the Enterprise product allows invoke to continue developing great software for everyone and release their studio for free as an open-source product.
What are the three main components of the invoke user interface?
-The three main components of the invoke user interface are the options panel, the work area, and the gallery.
What functionality does the work area provide in invoke?
-The work area allows users to review all of the images that have been generated in the studio, toggle between different generations, access additional information, and control a full suite of image actions.
What is the purpose of the boards and images/assets in the invoke gallery?
-The boards store images, which are the generations produced in the studio, or assets, which are images uploaded for use in control nets or IP adapters.
How does the 'text to image' feature in invoke differ from other image generators?
-In invoke, the terminology that users utilize is exactly what is passed into the model, providing full control and assurance that the prompt isn't being augmented with misaligned terms.
What are the two prompt boxes available in the 'text to image' tab of invoke?
-The two prompt boxes available are the positive prompt and the negative prompt, which can be used to control the image generation process.
What is the function of the dynamic prompts preview in invoke?
-The dynamic prompts preview allows users to see any dynamic props they're using to generate an image, providing a clear understanding of how the system processes each prompt.
How can the 'image accordion' in invoke be used?
-The image accordion is where users control the majority of the settings related to image size and starting noise, optimizing these based on the selected model architecture.
What is the workflow editor in invoke and how can it be utilized?
-The workflow editor allows users to use predefined and customizable workflows for generating content within the system. Users can save their workflows to the local database or use the default ones provided with the system.
What is the significance of the tiled upscaling workflow in invoke?
-The tiled upscaling workflow allows users to mimic the results seen in online tools like magnific ca, directly in invoke, providing a high degree of control over the upscaling process and the ability to augment the workflow for even more detailed results.
Outlines
🚀 Introducing Invoke: New Look and Enterprise Focus
The video script introduces the new look and name of Invoke, formerly known as Invoke AI. The company has shifted focus to include an Enterprise product aimed at larger creative teams in game studios, agencies, and media & entertainment companies. The script emphasizes that the support from these enterprises allows Invoke to continue developing software for everyone and to release their studio for free as an open-source product. The video will cover getting started with Invoke, the new user interface, and the basic use cases for new users. The interface is divided into options panel, work area, and gallery, each with specific functionalities and controls for image generation and manipulation.
🛠️ Enhanced User Interface and Generation Controls
The script delves into the specifics of the new user interface, highlighting the options panel where users can switch between modalities, update generation settings, and access powerful control capabilities. The work area allows review of generated images, toggle between generations, and control a suite of image actions. The gallery is split between boards and assets, used for storing generated images and uploaded images for control nets or IP adapters. The script also discusses the basic text-to-image generation tab, including the use of positive and negative prompts, embeddings, and dynamic prompts preview. The control of seed behavior and the image accordion for size and noise settings are also covered.
🎨 Advanced Features and Workflow Enhancements
This paragraph discusses advanced features such as control nut, IP adapter, and TOI adapter for controlling the composition of images. It also mentions the refiner tab for sdxl models and the advanced tab for less commonly used settings. The image-to-image tab updates are explained, including the use size button to match the initial image dimensions. The unified canvas is introduced as a tool for direct manipulation of image parts, with features like the scal before processing and coherence pass. The compositing section on the unified canvas is also detailed, including the coherence pass and infill tabs for additional matched details.
📚 Workflow Library and Additional Tips
The script introduces the workflow library feature, which allows users to save and access workflows for generating new content. Pre-installed workflows and the ability to augment them are discussed. The tiled upscaling workflow is highlighted as a way to mimic online tools for image upscaling with high control. The script also provides tips on using the workflow system, including the use of nodes and hotkeys for efficiency. The video ends with an invitation for professional creatives and community members to join Invoke and be a part of the exciting developments in AI for 2024.
Mindmap
Keywords
💡Invoke
💡Enterprise Product
💡User Interface
💡Text to Image Generation
💡Dynamic Prompts
💡Seed
💡Gallery
💡Control Net
💡IP Adapter
💡Unified Canvas
💡Workflow Editor
💡Tiled Upscaling
Highlights
Invoke AI is now rebranded as Invoke, continuing to build open-source software for creatives.
Invoke has released an Enterprise product aimed at larger creative teams in game studios, agencies, and media & entertainment companies.
The new user interface is designed to evolve over time, with AI integrated directly into the studio.
The options panel allows users to switch between modalities, update generation settings, and access powerful control capabilities.
The work area enables review of generated images, toggling between different generations, and accessing image information.
Users can control a full suite of image actions, such as opening in new tabs, downloading, and using image-to-image functions.
The gallery is divided into boards and images/assets, allowing for organization and use of images for control nets or IP adapters.
The text-to-image generation tab includes positive and negative prompts, embeddings, and dynamic prompts preview for enhanced control.
Dynamic prompts allow for generating images based on multiple conditions, such as 'red, green, or blue'.
The image accordion controls settings related to image size and starting noise, optimizing based on the selected model.
Advanced options are expandable and include settings like seed, high-resolution fix, and denoising strength.
The concepts tab showcases loaded luras, which can introduce new concepts directly into images.
Control section features like control nut and IP adapter enable users to control the composition of images using input images.
The refiner tab is for users who utilize the refiner model, offering controls for fine-tuning generation settings.
Seamless tiling in the advanced settings allows for generating assets with blended edges for textures and materials.
The image-to-image tab includes a use size button for matching the initial image dimensions and ensuring proper output shape.
Workflow editor and library have been introduced, allowing users to save and load workflows for content generation.
Tiled upscaling workflow mimics online tools like magnific ca, offering a high degree of control for upscaling images.