This is a workflow that utilises Deforum Nodes in ComfyUI to create various images that convey a somewhat linear animation for the use of videos. I currently use this workflow to create backgrounds on my music channel on <a target="_blank" rel="ugc" href="http://www.youtube.com/ChaosPromotions">YouTube</a>I have a discord if you would like to join to discuss AI, the workflow, music, play games or whatever else you can join here; <a target="_blank" rel="ugc" href="https://discord.gg/U5Ep2BTsfr">https://discord.gg/U5Ep2BTsfr</a> This workflow can be used in conjunction with the following tools which you may find useful;<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/articles/14641">https://civarchive.com/articles/14641</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/613826/sdxl-imagepath2vid-output-upscale-workflow">https://civarchive.com/models/613826/sdxl-imagepath2vid-output-upscale-workflow</a> </li></ul> Changelog: VERSION 14.0:<ul><li>ComfyUI Desktop Migration Version.</li></ul>VERSION 13.0:<ul><li>Small August Update, small changes, nothing important, just an update.</li></ul>VERSION 12.0:<ul><li>Added a Custom Strength Schedule as explained here; <a target="_blank" rel="ugc" href="https://civarchive.com/articles/14641">https://civarchive.com/articles/14641</a></li></ul>VERSION 11.0: Few Variation additions with small notes for guidance.<ul><li>Expanded Interpolation to 6 total frames instead of 4.</li><li>Changed Hybrid Motion to Farneback for quality gain in interpolation</li></ul>VERSION 10.0:<ul><li>DPM++ 2M Karras</li><li>ZavyChroma v.10</li><li>Optimised for personal use.(TRANSLATE IS BROKEN on ALL versions. IF ANYONE KNOWS HOW TO FIX PLEASE SEND ME A MESSAGE)</li></ul>VERSION 8.0:<ul><li>SDXL 1.0</li><li>Deforum Video Creation Workflow.</li><li>DDPM</li></ul>VERSION 5.0 Onwards:<ul><li>SDXL 1.0</li><li>Deforum Video Creation Workflow.</li><li>DPM++ SDE Karras</li><li>Perfect for use with ZavyChromaXL Checkpoint.</li><li>Strength is low due to reduce flickering.</li><li>Optimum CFG is 5-7 High detail, high depth.</li><li>Added -2 Clip Skip</li><li>Added VAE loader</li><li>Added Upscale Model Loader</li><li>Added Lora Capability</li><li>Refined Upscaler to use Ultimate SD Upscaler instead of default lanczos. Upscaler for 1920x1080 resolution (Can be bumped up to around 4k)Previous Versions have notes and the version number will have direct changes beside it) (Details may diminish over time with low Strength, recommend bumping to: 0.25-0.49 if you see detail disappear.) (When starting generation disable the Upscale node, Preview Image node and Save Image and Save Video node until one frame has been generated for the cache to pull from, after first generation enable all above nodes and all should be good!)If you want to save frames, use Save Image Extended, if you want to create video use Save Video. Working on this for the past 3 months. Finally got to a place I'm happy. Flickering is still present as Deforum is temperamental, until I find a different selection of nodes to achieve the same thing this is probably the best I can achieve with Deforum. But who knows, there may be future iterations of this workflow. Thanks for all the support on this journey and feel free to use and post anything you create with this workflow alongside the Gallery of this submission. Appreciate all the buzz and support and downloads. Thankyou! ♥</li></ul>This Deforum ComfyUI workflow is my test for generating backgrounds for YouTube videos over <a target="_blank" rel="ugc" href="http://www.youtube.com/ChaosPromotions">here</a>.

JAFFA's Custom Deforum Workflow

Ready to go. Optimised for 720p generations with Optical Flow generation. Always welcome to experiment yourself and do whatever you like. Thanks for over 150 downloads ♥ May keep updating this on a per-video basis as time goes on. Still tweaking and finding the sweet spot for certain things.CALL TO ACTION: Also I would like to take advantage of the built in Deforum Comfy Nodes related to Audio and Amplitude, If anyone knows any informative videos or tutorials or workflows using the amplitude features of deforum in ComfyUI please let me know I'd appreciate it a lot as I'd love to implement this and take advantage of those features if they're readily accessible. Feel free to leave comments pointing me in directions that you might find useful for this. It would save me a lot of work to do things inside ComfyUI instead of having to use external software to achieve results. Thanks for reading and downloading I'd love to hear your feedback where this workflow can be improved or adapted to be more accessible/optimised! 🔥♥❤️♥🔥-jaf<a rel="ugc" href="http://www.youtube.com/ChaosPromotions">Chaos</a>

Files with SHA256: 7966af686a6f0a393537121a252f2516f5b5f44c0baf8856859c9b38bbabeded