Innovate Futures @ Benji posts
Posts: 709
Files: 875
(Workflow Update) Wan2.2 Animate (native) (Swap Multi-Person Video) (Long Length Loop) (DetailEnhancer) Ver 20251027 (Patreon Supporter)
Previous Versions: Ver 20251017Added Continue Generation Video.If you enable this group, you can input a generated video, and continue with the last frames.This help Wan 2.2 Animate due to memory limitation, even in theory it can generate unlimited long length by Loop. And AI model degradation, as soon as over 15 seconds, color keep degrading.So for running over 1 or 2 minutes long video, separate by batches and continue from the last video fram...
MoCha New Fine-Tuned Video Model - Easier Than Wan 2.2 Animate And It Works!
Video : https://youtu.be/yypmE5VXkroIn this video, we explore MoCha New Fine-Tuned Video Model based on WAN 2.1 that enables high-quality character replacement in videos using just a single reference frame. Unlike WAN 2.2 Animate, Mocha doesn’t require complex ControlNet setups or multi-frame masking, making it faster and easier to use while still delivering impressive results with dynamic lighting, facial fidelity, and motion tracking. This tu...
Ditto in ComfyUI With Wan Vace - Transform ANY Video with Just a Text Prompt!
Video : https://youtu.be/yvIubLbSJ-gIn this video, you’ll learn how to easily install and use the official Ditto models in ComfyUI to restyle your videos using only text prompts—no reference images needed. We walk through both the official GitHub workflow and native ComfyUI node setups, showing you how to load the Ditto Global Comfy model, configure LoRAs like COS V14B, and generate stunning style transfers while preserving original motion. T...
ComfyUI Qwen 3 VL Create Powerful Prompts For Image And Video Generation
Video : https://youtu.be/3j9c_-mRKfgIn this video, you’ll learn how to use ComfyUI Qwen 3 VL—a powerful vision language model—directly inside ComfyUI to generate detailed text prompts from images or videos, and then use those prompts to create new AI-generated content. We walk through both image and video workflows, showing how Qwen VL can analyze visual input and produce rich, time-coded descriptions that feed into diffusion models like WA...
FlashVSR UpRes With 1 Step - Explore This AI Video Enhancer With Other Methods in ComfyUI!
Video : https://youtu.be/3EA-PzOxk-0In this video, we dive into the Flash VSR AI video upscaling model and its integration within the ComfyUI Wan Video Wrapper. The creator explores how Flash VSR enhances low-quality or compressed AI-generated videos by increasing resolution, improving frame interpolation, and reducing artifacts like blur and pixelation. While the model is now accessible through custom nodes in ComfyUI, the video highlights a cri...
[Workflow Update] Wan2.2_Animate (native) (Swap Person In Multi-Person Video) (Long Length Unlimited Loop) (DetailEnhancer) Ver 20251017
Previous Version : Wan2.2_Animate (native) (Swap Person In Multi-Person Video) (Long Length Unlimited Loop) (DetailEnhancer) Ver 20251009https://www.patreon.com/posts/workflow-update-140836677Wan 2.2 Animate (Long Length Gen) Ver.20250926 - Updatehttps://www.patreon.com/posts/wan-2-2-animate-139775703Ver 20251017 update:Try with new Segmentation model, instead of SAM 2, let's try Sec4B for segmentation.About Sec 4B Segmentation:https://www.pa...
Sec 4B New Way For Video Mask Editing - SAM2 Is Obsolete?
Video : https://youtu.be/zByo6_W9FN8In this video, we walk through how to install and use the new SAG (Stack Anything Generalized) video segmentation model in ComfyUI—comparing its performance side-by-side with SAM2 for real-world video editing tasks. You’ll see how SAG handles complex scenes with fast motion, camera cuts, and partial occlusions, and how to integrate it into workflows like WAN 2.2 Animate for advanced AI video generation. Thi...
[update]OVI Video X Qwen Image Edit - Next Scenes
Related Post: https://www.patreon.com/posts/qwen-image-edit-141282221For Qwen Image Edit AI video scenes creation.There are an alternative way to create image2video from Qwen Image.We have talk about Ovi AI Video before, and I edit that OVI multi scenes workflow for this Qwen Image Edit - Next Scenes workflow as well.So you can use the Qwen Image Edit workflow in the public post , when the OVI video.If you have any question, please feel free to l...
Qwen Image Edit & Wan 2.2 - Create Coherent AI Video Scenes With This!
Tutorial Video : https://youtu.be/YQLq--X--HYIn this video, we explore a powerful AI storytelling pipeline that combines language models, text-to-image generation, and image-to-video workflows to create structured, multi-scene AI videos. Instead of relying on a single reference image or generating random clips, the creator demonstrates how to use Qwen 3 Max to generate a sequence of detailed text prompts—each describing a specific scene with su...
[update] Wan Video I2V - InfiniteTalk - Long Content Talking Avatar Generator (Single) (Ver.20251014)
previous version: Ver.20250822 https://www.patreon.com/posts/for-patreon-wan-137035392Other related post:V2V - InfiniteTalk - Long Ver.20250823https://www.patreon.com/posts/for-patreon-wan-137287049InfiniteTalk V2V Test Exampleshttps://www.patreon.com/posts/infinitetalk-v2v-137187389Updateupdate Chatterbox node
Wan Vace With Lynx And Infinite Talk - Create the Most Realistic Talking Avatar Video Ever!
Tutorial Video : https://youtu.be/wQ6zf9wxo7QIn this video, we dive deep into the advanced capabilities of Wan VACE combined with Lynx Face ID and Infinite Talk, showcasing how to create hyper-realistic talking avatar videos using ComfyUI. Whether you're a content creator, filmmaker, or marketer, this tutorial will walk you through the step-by-step process of achieving seamless face swaps, consistent character identities, and natural lip-sync...
Wan Lynx AI Video Model In ComfyUI With Vace! This AI Creates High-Fidelity Video Generation!
Video: https://youtu.be/TJeNdpC_GpYDiscover ByteDance Lynx, the latest breakthrough in high-fidelity personalized video generation—powered by the Wan 2.1 foundation model and now fully compatible with ComfyUI! In this deep dive, we explore how to run Lynx locally using the Wan Video Wrapper, including the key components: ID adapter, reference adapter, and resampler. This setup enables you to generate text-to-video clips with consistent facial i...
Ovi AI Video in ComfyUI Q&A/ Support Area Here (Patreon Supporter)
Video: https://youtu.be/6SNgxL6mKnQAny installation, model download need help, free feel to ask in the comment section below.Ovi: Twin Backbone Cross-Modal Fusion for Audio-Video Generationhttps://huggingface.co/chetwinlow1/Ovihttps://github.com/character-ai/Ovihttps://huggingface.co/papers/2510.01284ComfyUI Custom Node: ComfyUI-Ovihttps://github.com/snicolast/ComfyUI-OviDownload Weights via https://github.com/snicolast/ComfyUI-Ovi?tab=readme-ov-...
Ovi AI Video - Veo 3 Alike Video In ComfyUI (Tutorial)
Video : https://youtu.be/6SNgxL6mKnQResourcesOvi: Twin Backbone Cross-Modal Fusion for Audio-Video Generationhttps://huggingface.co/chetwinlow1/Ovihttps://github.com/character-ai/Ovihttps://huggingface.co/papers/2510.01284ComfyUI-Ovihttps://github.com/snicolast/ComfyUI-OviDownload Weightshttps://github.com/snicolast/ComfyUI-Ovi?tab=readme-ov-file#weights-overviewMMAudio/ext_weights/{best_netG.pt, v1-16.pth}https://huggingface.co/hkchengrex/MMAudi...
Wan 2.2 Animate V2 Update And Video Detail Enhancer - This Trick Fixes Blurry Hands & Faces
Video : https://youtu.be/pwA44IRI9tAWan 2.2 250928 Lightning https://youtu.be/tIQNxi95AzQWan Animate Native Node https://youtu.be/W7-u2tYZiFoIn this video, we dive deep into the WAN 2.2 Animate V2 update—a major upgrade that significantly improves motion quality, facial expression accuracy, and overall video coherence. You’ll see how the new model handles complex scenes like dynamic camera angles and outfit details (like T-shirts, jackets, an...
Wan 2.2 New Lightning and Fine-Tune Models - It Fixed All the Old Problems!
Tutorial Video : https://youtu.be/tIQNxi95AzQ?si=OXKf8nu-4R_WSL7EIn this video, we dive deep into the latest updates for WAN 2.2 AI video generation—specifically the new Lightning models and fine-tuned variants like WAN 2.2 Palingenesis. You’ll see real-world comparisons between older WAN 2.2 Lightning versions and the newest high-noise diffusion model that merges Light X2V technology for dramatically improved motion, camera dynamics, and ge...
How To Use Wan 2.2 And Qwen Image Models Pipeline - Alibaba AI Ecosystem For Content Creation
Video : https://youtu.be/qMda40fB_m0In this video, we break down a complete local AI content creation pipeline using Alibaba’s latest open-source models: WAN 2.2 Animate for AI video animation and Qwen Image for high-quality image generation and editing.You’ll see how to connect these tools inside ComfyUI—from generating a scenes and character with Qwen Image, image editing using Qwen Image Edit 2509, to animating the scene with realistic m...
[Update] Wan2.2 Fun Vace (Native) (ControlNet V2V) (Long Video Gen.) Ver.20250927
Related Posts:Wan2.2 Vace Fun StartEndFrame LongVideoGen All-In-Looping Ver202509(Update) Native Wan VACE Outpaint - Long Video Gen - Version 20250911[For Patreon Supporters] [Workflow Update] Wan_Vace-Native-V2V-CN LongVideo (Unlimited Length) 20250813[For Patreon Supporters] [Workflow Update ]Wan Vace 2.1 - Image2Video LongLengthGen - Ver.20250807Vace ControlNet V2V Long Length Gen. This is another workflow update for Vace. This workflow are us...
Wan 2.2 Animate (Long Length Gen Looping (Advanced Pose) Ver.20250924) - Additional For Patreon Supporters
In the previous public post, I have post the general Long Length Gen Looping method for Wan 2.2 Animate native node. https://www.patreon.com/posts/139637572In this post, I am going to post additional enhancement for this workflow.1 - DwPose ExtractIf you have low VRam , or you are processing very long length ,e.g: over 2 minutes duration.I suggest to have another workflow that only take care of extracting DWPose and Face video first.It helps a lo...
Wan 2.2 Animate Support In ComfyUI Native - TikTok Dance Videos Go to Next Level!
Tutorial Video : https://youtu.be/W7-u2tYZiFoIn this video, we dive deep into running WAN 2.2 Animate natively in ComfyUI—no wrappers, no extra tools, just pure node-based workflow power. You’ll see how to generate high-quality character-swapped animations and pose transfers directly inside ComfyUI, using the latest updates including support for long-form video generation, frame offset chaining, and efficient background handling. We’ll walk...
SRPO txt2img & img2img Refiner Upscale Ver.20250923 (For Patreon Supporters)
Related Post: https://www.patreon.com/posts/139541646Tutorial Video : https://youtu.be/qEU19uWk5ukAs mentioned in this tutorial, attached all workflows including the freebies one here.Image Gen. pretty easy nowadays, just depends on the AI model quality.But still leave a comment if you have any questions.
Flux Hunyuan SRPO - Stop Using Flux Alone — SRPO Makes It 10x Better (Tutorial Inside)
Video : https://youtu.be/qEU19uWk5ukIn this video, I’ll walk you through everything you need to know about downloading, installing, and using the Flux Hunyuan SRPO model — whether you’re grabbing the full 47GB version or the lightweight GGUF quantized files perfect for your home PC. I’ll show you exactly where to find the files on Hugging Face, how to load them into ComfyUI using the custom GGUF node, and how to set up your folder structu...
Wan 2.2 Animate In ComfyUI Using WanVideoWrapper - Tutorial And Usage Examples
Tutorial Video : https://youtu.be/lXS5ZdCcnuoUnlock the full power of Wan 2.2 Animate In ComfyUI Using WanVideoWrapper! This step-by-step tutorial walks you through setting up the Wan Video Wrapper, loading the correct models (FP8, GGUF, or BF16), and mastering the new “Wan Videos Animate” node to create stunning, customizable character animations. Learn how to swap characters, control poses, match facial expressions, and even handle complex ...
Wan 2.2 Animate A Groundbreaking Model & VACE Fun Long-Length Workflow!
Video : https://youtu.be/6cDWP5AY7CQIn this video, we dive deep into the brand-new Wan 2.2 Animate and Vase Fun models — the latest breakthroughs in AI video generation. You’ll see exactly how to swap characters, create ultra-smooth long-form animations using start/end frames, and fix common issues like reversed motion or color shifting. I’ll walk you through my custom ComfyUI workflows, show you how to use overlap frames for seamless stitc...
Wan2.2 Vace Fun StartEndFrame LongVideoGen All-In-Looping Ver202509 (Patreon Supporters Q&A)
Video : https://youtu.be/6cDWP5AY7CQRelated post : https://www.patreon.com/posts/139254327Wan2.2 Vace Fun , flip back to the previous tutorial if you don't know what is this : https://youtu.be/hTZfTj5wIiYResources:alibaba-pai/Wan2.2-VACE-Fun-A14Bhttps://huggingface.co/alibaba-pai/Wan2.2-VACE-Fun-A14BQuantStack/Wan2.2-VACE-Fun-A14B-GGUFhttps://huggingface.co/QuantStack/Wan2.2-VACE-Fun-A14B-GGUFComfy-Org/Wan_2.2_ComfyUI_Repackagedhttps://huggin...
HuMo ByteDance’s New Reference2Video With Audio Lipsync - Keep Or Toss?
Video : https://youtu.be/DB7wGLFp4LQAI video generation with HuMo — the new human-centric model from ByteDance and Tsinghua University. In this hands-on tutorial, I’ll walk you through how to install, configure, and generate talking character videos using HuMo in ComfyUI. Whether you’re feeding it text + audio, text + image, or all three together, you’ll learn how to create cinematic, lip-synced avatars with realistic expressions — all ...
Wan2.2 VACE Fun Released! Tutorial Guide Update To Your VACE Workflows
Video : https://youtu.be/hTZfTj5wIiYHey everyone, welcome back! In this video, I’m diving deep into the brand-new Wan2.2 Fun VACE model — the latest experimental release from Alibaba PAL that’s already shaking up the AI video community. If you’ve been using Wan 2.1 VACE to create long-form videos or fine-tune character consistency, this is the upgrade you’ve been waiting for.I’ll walk you through exactly how to download, load, and run...
MimiCPM-V 4.5 Vision LM - Ran GPT-4o-Level Vision AI Locally Or Handheld
Tutorial Video : https://youtu.be/oSqj9DuXK5IIf you’ve ever wanted to run GPT-4-level vision AI right on your laptop—or even your iPhone—without uploading a single photo to the cloud… this video is for you. I’m diving deep into MiniCPM-V 4.5, the open-source, 8-billion-parameter model that’s beating GPT-4o and Gemini 2.0 Pro on image and video understanding—all while running locally, offline, with no internet needed.In this walkthro...
(Update) Native Wan VACE Outpaint - Long Video Gen - Version 20250911
Native Wan VACE Outpaint - Long Video Gen - Version 20250911Update & Fixed:Looping count and conditionLocal Ollama Set as optionalCurrently, I like to use FusionX Vace load as Lora model. It's more flexible for my setup, if you have FuxionX Vace as GGUF or Diff. Model, you can use yours and remove the FusionX Vace Lora.Each 129 frames or 81 frames sampler frames should be follow the next segment of the whole video. It have not skip few fram...
Wan2.2 Reward LoRAs MPS & HPS TestLab workflow And More In-Depth.
Tutorial Video : https://youtu.be/2xpOCCTeSXoRelated Post: https://www.patreon.com/posts/138580645About HPS And MPShttps://github.com/tgxs002/HPSv2https://github.com/Kwai-Kolors/MPSAlibaba-pai/Wan2.2-Fun-Reward-LoRAshttps://huggingface.co/alibaba-pai/Wan2.2-Fun-Reward-LoRAs(Download Into your models/loras/ folder)Here's a clear breakdown of your questions, focusing on technical distinctions and practical implications:1. Advantage of Reward Lo...
