The GitHub repository integrates Gemini into ComfyUI, offering models like Gemini-pro, Genimi-pro-vision, and Gemini 1.5 Pro for text, image, and file processing tasks. Users can apply for their own API Key to access Gemini API. The repository provides workflow examples, installation instructions, and updates. Contact information includes zhozho3965@gmail.com and a QQ Group (839821928). Social media links to ‘-Zho-’, Bilibili, Twitter, Little Red Book, and support on Bilibili and Aifadian. Credits to ComfyUI_Custom_Nodes_AlekPet. - Users need to apply for a Gemini_API_Key to use Gemini nodes, ensure a stable connection to Google Gemini’s services, and update the dependency ‘google-generativeai’ to version 0.4.1 for Gemini 1.5 Pro. Installation can be done using ComfyUI Manager or manually by cloning the repository from GitHub and installing requirements. Workflow versions include V3.0 with Gemini 1.5 Pro, V2.0 with a chatbot workflow, and V1.1 with workflows for Gemini-pro and Genimi-pro-vision. Updates include Version 3.0 adding Gemini 1.5 Pro, system instructions, and file uploads, Version 2.1 fixing a bug, and Version 2.0 adding context chat nodes. - The status history for the repository is available.
The video titled ‘Expanding Horizons: Outpainting Mastery in ComfyUI’ explores the artistry of Outpainting with ComfyUI’s Stable Diffusion feature. The video delves into Hyper Expansion, Sketch to Render, and Auto Background Regeneration within the realm of Outpainting. Topics discussed include the use of LORAs, Quick Basic Outpainting, and various techniques for Outpainting with ComfyUI.
The Aerial view of the building showcases a LoRA DoRA etc model for urban bird’s-eye views, offering high-definition training sets for cityscapes and buildings. The model, based on SD 1.5, has received very positive reviews and was last updated on Aug 3, 2023.
The GitHub repository showcases ComfyUI, a powerful and modular Stable Diffusion GUI, API, and backend with a graph/nodes interface. The interface allows users to design and execute advanced stable diffusion pipelines without needing to code. ComfyUI supports various features like SD1.x, SD2.x, SDXL, Stable Video Diffusion, and Stable Cascade. Users can experiment with complex workflows, embeddings/textual inversion, Loras, hypernetworks, and more. The repository provides detailed installation instructions for Windows, Linux, AMD GPUs, NVIDIA GPUs, Intel GPUs, Apple Mac silicon, and DirectML for AMD cards on Windows. Additionally, it offers shortcuts for workflow management, high-quality previews, TLS/SSL setup, and support channels for users. The repository is licensed under GPL-3.0 and has garnered 35.3k stars and 3.8k forks.
The video titled ‘How 2 Canvas Node’ is available on YouTube. The video duration is 14 minutes and 22 seconds.
The video showcases a plugin that provides realtime AI assistance to krita, a digital painting software. The channel, Nerdy Rodent, offers tutorials on Stable Diffusion, Generative AI, Large language models, and other AI tools, catering to AI enthusiasts and professionals. The plugin enhances various aspects of AI technology, such as voice cloning, text-to-speech, and style transfer, making it a valuable resource for AI enthusiasts and artists alike.
The video titled ‘ComfyUI Fundamentals - Upscaling 1’ explores the concept of upscaling in the context of ComfyUI. The video delves into how upscaling works and provides tips to enhance the upscaling process. The presenter discusses the ability to go under 1 on the upscale by node, highlighting key aspects of image upscaling and various upscaling options.
The YouTube video titled ‘ComfyUI Modular- Ultimate Starter Workflow Usage’ provides insights into using ComfyUI Modular for an ultimate starter workflow.
The YouTube page provides information about the use of cookies and data for various purposes, including Google services, audience engagement, and site statistics. By accepting all, users allow for the development of new services, displaying ads, and showing personalised content and ads based on their settings. Non-personalised content and ads are influenced by factors like the viewed content and location. Users can manage privacy settings and find more information by selecting ‘Meer opties’.
The GitHub repository is an attempt to use TensorRT with ComfyUI, focusing on specific models and workflows. The repository provides instructions for installation and usage, highlighting compatible models like Stable Diffusion and SDXL. It also mentions limitations and future improvements for user-friendliness and compatibility with ComfyUI-Manager. The repository is a ComfyUI port from the official A1111 extension, with potential maintenance issues addressed by an alternative repository. The README outlines conversion scripts, dependencies, and error messages encountered during the process.
The GitHub repository showcases experimental usage of stable-fast and TensorRT. It includes speed tests, installation instructions, and features of stable-fast and TensorRT.
The repository provides guidance on enabling stable-fast nodes and installing TensorRT for testing purposes.
It highlights the compatibility of stable-fast with LoRA DoRA etc, ControlNet, and other models, along with speed optimizations and node support.
Speed tests on a GeForce RTX 3060 Mobile show performance metrics for stable-fast and TensorRT implementations.
The repository also includes a table detailing features, tested nodes, and performance benchmarks for different workflows.
The GitHub repository provides a streamlined interface for generating images with AI in Krita. Users can inpaint and outpaint with optional text prompts without the need for tweaking. The plugin allows for creating new images from scratch, refining existing content, and live painting. It supports various resolutions, job queues, and history tracking. Customization options are available for advanced users. The technology used includes Stable Diffusion for image generation, ComfyUI for the diffusion backend, ControlNet for Inpainting, and IP-Adapter for outpainting. The repository is licensed under GPL-3.0 and has garnered 4.8k stars and 216 forks. Contributors have added features like object selection tools and GPU cloud support.
The ComfyUImodels tag on Civitai features 379 models. Civitai offers a range of AI models for various applications and purposes.
The GitHub repository titled ‘sharing-is-caring’ is maintained by the fictions.ai team. They focus on collaboration and sharing knowledge related to A.I. generation. The repository contains various workflows, scripts, and tools for A.I. generation. Some key contents include Comfy Workflows, upscale workflows, and specific workflow requirements. Contact the fictions.ai team for questions or feedback.
The Index page of ComfyUI Resources provides a comprehensive list of custom nodes and tools for ComfyUI. Various nodes such as CLIP BLIP Node, GPT node ComfyUI, and Vid2vid Node Suite are available for installation. Instructions for installing custom nodes are included on the page.
The Comfy Workflows website offers a platform for sharing art and workflows, with a focus on ComfyUI. Users can explore thousands of workflows created by the community and run them with zero setup using the ComfyUI Launcher. The site features a variety of workflows, images, and videos created by different creators.
The GitHub repository integrates Gemini into ComfyUI, enabling users to generate prompts, describe images, and converse with Gemini. The repository features the latest Gemini 1.5 Pro model with system instruction settings, multi-modal conversations, and file reading capabilities. Users can request their API Key for Gemini API. Various models and nodes are provided for different functionalities, along with installation methods, workflows, changelog, and contact details.
The video titled ‘Expanding Horizons: Outpainting Mastery in ComfyUI’ on YouTube showcases the mastery of outpainting in ComfyUI. The video explores the revolutionary design aspects, including Vignette Mastery, ComfyUI Magic, and Font Previews Galore.
The paper titled FreeU: Free Lunch in Diffusion U-Net explores the potential of enhancing generation quality in Diffusion Models without additional training. The authors propose a method called ‘FreeU’ that strategically re-weights contributions from U-Net’s skip connections and backbone feature maps to improve generation quality. The results show promising outcomes for image and video generation tasks, demonstrating the ease of integration with existing diffusion models.
The GitHub repository explores the attempt to use TensorRT with ComfyUI, focusing on compatibility and installation instructions. The repository provides information on supported models like Stable Diffusion and SDXL, along with a list of working and non-working models. The project aims to make the process more user-friendly and automatic in the future. The README.md file outlines the steps for installing Python dependencies, TensorRT versions, and converting Checkpoints to TensorRT engines. It also discusses the usage of converted engines in ComfyUI and common error messages. The repository includes links to original implementations and download links for testing with various checkpoints and models.
The AP Workflow 9.0 for ComfyUI introduces new features such as upscalers, image generation with Dall-e, advanced XYZ plot, face cloner, face analyzer, and training helper for batch captioning. Instructions for troubleshooting custom node installation, switching to SD 1.5 models, using LM Studio for prompt enrichment, securing ComfyUI connection with SSL, and FAQs are provided. Special thanks to custom node creators and a full changelog for version 9.0 are included. The AP Workflow version 8.0 offers features like bookmark nodes, IPAdapter plus v2 nodes, uploader function, caption generator function, image evaluators, face analyzer, aesthetic score predictor, image chooser, prompt enricher function, LoRA DoRA etc info node, face detailer function, and reorganized L2 pipeline layout with removed functions like ReVision and Image Enhancer.
The GoogleSpreadsheets document titled ‘SDXL Model Compare’ provides information on various data models and comparison metrics. The content includes data on different styles, illustrations, and financial aspects. The document also offers support for screen readers and data cleaning suggestions.
The GitHub repository for ComfyUI-MotionCtrl contains an implementation of MotionCtrl for video generation. The repository includes nodes for loading motion control Checkpoints, motion control conditioning, and motion control Sampling. Tools for generating motion trajectories and camera points are also provided. Examples of workflows for generating LVDM/VideoCrafter videos and using AnimateDiff for scribbling are available. The repository is licensed under Apache-2.0 and has received 121 stars and 4 forks.
The YouTube page provides information about cookies and data usage for Google services. Users can choose to accept or reject cookies for various purposes, including delivering Google services, tracking outages, measuring audience engagement, and showing personalised content and ads based on user settings.
The video titled ‘How 2 Canvas Node’ is available on YouTube.
The YouTube page provides information about cookies and data usage by Google services. Users can choose to accept or reject cookies for various purposes, including delivering Google services, measuring audience engagement, and showing personalised content and ads. More options are available for managing privacy settings.
The GitHub repository provides a streamlined interface for generating images with AI in Krita. It allows users to inpaint and outpaint images with optional text prompts, requiring no tweaking. The plugin supports features like generating images from scratch, refining existing content, live painting, and job queue management. Customization options are available for advanced users. The plugin is open source and free to use. krita-plugin, stable-diffusion, Generative AI
The YouTube video showcases a plugin that adds real-time AI assistance to krita. The video covers topics such as Stable Diffusion, Generative AI, Large language models, AI Animation, Voice Cloning, and more. The channel, Nerdy Rodent, provides tutorials on Artificial Intelligence in an easy-to-digest format. The AI enthusiast behind the channel recommends specific hardware for the best AI experience at home. Not suitable for children. Please use AI responsibly.
The GitHub repository showcases ComfyUI, a powerful and modular Stable Diffusion GUI, API, and backend with a graph/nodes interface. The interface allows users to design and execute advanced stable diffusion pipelines without needing to code. ComfyUI supports various Diffusion Models, asynchronous queue system, and many optimizations to enhance workflow efficiency. Users can experiment with complex workflows, including area composition, Inpainting, controlnet, upscale models, and more. The repository provides detailed installation instructions for Windows, Linux, AMD GPUs, NVIDIA GPUs, Intel GPUs, Apple Mac silicon, and DirectML for AMD cards on Windows. Additionally, it offers shortcuts for workflow management, high-quality preview setup, TLS/SSL configuration, and support channels for users. The repository is licensed under GPL-3.0 and has garnered significant community engagement with 35.3k stars and 3.8k forks.
The video titled ‘LATENT Tricks - Amazing ways to use ComfyUI’ features Olivio Sarikas, an AI Expert and passionate Artist, showcasing the exciting world of AI art. The video invites viewers to explore AI art and creative visions with live streams. Olivio Sarikas, a professional Designer with a Masters Degree in Fine Arts, shares Tips and Tricks for using ComfyUI.
The ComfyUImodels tag on Civitai features 379 models for Stable Diffusion AI. Users can explore and access these models for various applications. Civitai offers a range of services and resources for creators, including terms of service, privacy policies, and safety guidelines.
The ComfyUI-extension-tutorials/ComfyUI-Impact-Pack/workflow on GitHub provides a collection of workflow-related files for various image processing tasks. The repository includes files for tasks such as Upscaling, animation, segmentation, and more. Each file represents a specific workflow step or technique, showcasing the versatility and capabilities of the ComfyUI Impact Pack.
The video titled ‘ComfyUI Impact Pack - Q&A;: Detailer Options’ on YouTube provides explanations on the important parameters of the detailer.
The Index page of ComfyUI Resources provides a comprehensive list of custom nodes and tools available for use. The nodes cover a wide range of functionalities, from image processing to AI installation tools. Installation instructions can be found on the respective node pages, with the option to streamline the process using ltdrdata’s Comfy Manager.
The YouTube video titled ‘ComfyUI Fundamentals - Upscaling 1’ provides insights into the fundamentals of ComfyUI and the concept of upscaling in the context of user interface design.
The YouTube video titled ‘ComfyUI Modular- Ultimate Starter Workflow Usage’ provides insights into using ComfyUI Modular for an ultimate starter workflow.
The GitHub repository for Prompt-Free Diffusion discusses a diffusion model that generates images using only visual inputs, replacing text encoders with a Semantic Context Encoder (SeeCoder). The model is reusable across various T2I models and adaptive layers. The repository includes pretrained models and tools for model conversion. The implementation is based on a research paper presented at arXiv 2023 / CVPR 2024.
The file contains Python code for a custom node in the ComfyUI/custom_nodes directory. The code defines a class for handling model references and latent data. The JSON file included in the gist outlines a workflow with various nodes like VAEDecode, CLIPTextEncode, KSampler, SaveImage, CheckpointLoaderSimple, ImageScale, LoadImage, and ReferenceOnlySimple.
The GitHub repository contains a Node.js WebSockets API client for ComfyUI. The client is based on the WebSockets API example and is licensed under the MIT license. The repository includes folders and files such as examples, source code, and configuration files. The client allows users to connect to a server, generate images based on prompts, and save the images to a specified directory. Topics related to the project include nodejs, api, stable-diffusion, comfyui, and sdxl.
The YouTube page provides information about cookies and data usage by Google services. Users can choose to accept or reject cookies for various purposes, including personalised content and ads. More options are available for managing privacy settings.
The video showcases ComyUI, a tool for video animation rendering using AI technologies like WAS, Seecoder, Style, and Semantic segmentation. The creator, Amir Ferdos, a seasoned 3D artist and designer, explores the fusion of AI with design processes, offering unique workflows and tutorials on their YouTube channel. The videos aim to educate and inspire designers on the transformative impact of AI in design. For a deeper dive into the creator’s work, exclusive tutorials and source files are available on their Patreon page.
The ComfyUICommunity Manual provides documentation for ComfyUI, a Stable Diffusion GUI and backend. It covers topics such as installation, downloading models, first steps with Comfy, loading other flows, and further support. The manual includes detailed information on interface, core nodes (including advanced, conditioning, experimental, image, latent, loaders, mask, and Sampling), custom nodes, developing custom nodes, and contributing documentation.
The GitHub repository provides a powerful tool that translates ComfyUI workflows into executable Python code. The tool bridges the gap between ComfyUI’s visual interface and Python’s programming environment, streamlining the process for data scientists, software developers, and AI enthusiasts. Use cases include creating lean app deployments, programmatic experiments, and large image generation queues. The v1.0.0 release notes highlight support for custom nodes. To use the tool, clone the repository, enable Dev mode options in ComfyUI, save workflows in API format, and run the script to generate Python code for image generation without launching a server. The repository is primarily focused on topics like pytorch, generative art, image generation, AI art, Stable Diffusion, and ComfyUI.
The website offers ComfyUI Cloud services for running and deploying workflows without the need for downloads or installs. Users can pay only for active GPU usage, avoiding idle time and unnecessary costs. ComfyICU provides ready-to-use creative workflows and a simple, scalable API for production. The platform aims to simplify workflow creation and deployment, offering fast performance and cost-efficiency. Users can access over 5000 happy users’ testimonials and FAQs for more information.
The video titled ‘EASY Inpainting in ComfyUI with SAM (segment Anything) | Creative Workflow Tutorial’ provides a tutorial on using SAM for inpainting in ComfyUI.
The wiki page provides a practical and collaborative guide on developing custom nodes for ComfyUI. The guide is unofficial and focuses on practicality over formality, encouraging collaboration through Q&A-style discussions. It covers various topics related to custom node development, such as control flow, data types, and UI design.
The Comfy Workflows website offers a platform for sharing art and workflows, with features like ComfyUI Launcher for running workflows with zero setup. Users can explore thousands of workflows created by the community. Trending creators and the latest images and videos are showcased on the site.
Custom nodes for interpolating between, well, everything in the Stable DiffusionComfyUI. The GitHub repository contains functionality to create preprocessed ControlNet OpenPose inputs midway between two images. Future features include line-art interpolation. To install, follow the provided instructions.
The GitHub repository contains ComfyUI Extension Nodes for Automated Text Generation. The repository is under development, with features like autogen, automated task solving, and group chat capabilities. The repository includes various folders and files for different functionalities. To contribute, users can submit pull requests, suggestions, or issue reports. The repository is licensed under AGPL-3.0. The repository has 314 stars, 24 forks, and 4 contributors.
The AutoGen Advanced Tutorial on YouTube focuses on building incredible AI AGENT teams. The tutorial delves into advanced techniques for creating AI teams and enhancing their capabilities.
The GitHub repository provides a TouchDesigner interface for ComfyUI, offering features like workflow creation and image send/receive. The repository includes installation instructions and resources for using the TDComfyUI component. It also offers guidance on connecting to Stable Diffusion and utilising ComfyUI settings for optimal performance.
The GitHub repository showcases the ComfyUI plugin for Photoshop, offering AI-powered image generation features. The plugin enables unlimited generative fill, customizable back-end workflow, and one-click image transformation. System requirements include a minimum of 6GB graphics memory and 12GB RAM. Installation involves downloading the plugin from a provided link or locally via a .CCX file. Additional files are required for specific functionalities. Support and contributions are encouraged through GitHub.
The model titled ‘Aerial view of the building’ offers a high-definition training set for urban bird’s-eye views, encompassing a variety of domestic and foreign architectural drawings. The model is based on LoRA DoRA etc technology and has received positive reviews. The training set is designed for cityscape and building enthusiasts.
