Innovate Futures @ Benji posts
Posts: 709
Files: 875
80 Generates for 1 scene... finally gamble a good Seed number.
This weekend, lots of fun. I never try with 80 times click click click... and generate a scene I want. LOL share this later.
Prep for another new AI Video :)
I try to apply all recent new tech together in the new AI video. I will share some thoughts and experience in Patreon afterwards. And see if any AI and workflows should be recommend.Should be fun. Off grid for few days, see ya.
Misto Line ControlNet For Flux 1 Dev In ComfyUI - The Most Detail ControlNet Model!
Video : https://youtu.be/LSBilDkqdyYIn this video, we delve into the intricacies of the Misto line for Flux, a powerful line trace preprocessor for ControlNet. We explore the detailed capabilities of the Misto line for Flux One Dev model, showcasing its compatibility with the GGUF model and FP16 models. Through side-by-side comparisons, we reveal the stunning results of utilizing Misto line preprocessors for generating AI images with exceptional detail and accuracy. Join us as we navigate through the installation process, custom nodes, and work...
Cursor AI For Beginners From Text Prompt To AI Coding
Video : https://youtu.be/nc8CinDcLOMIn this video, I dive deep into the realm of AI programming with a focus on the innovative tool called Cursor. Join me as I explore this AI training tool, breaking down its capabilities and functionalities. From generating code snippets to creating basic games like chess and snake, we uncover how Cursor simplifies the coding process for both beginners and seasoned programmers alike. Discover the potential of this tool in enhancing your backend programming tasks and navigating through AI-generated scripts with...
MimicMotion-V4_workflow - For Multi-Characters Animate
This is another update for Mimic Motion V4 workflow. The Version 4 series are going for multi-characters animation.In this demo, I am doing for 2 characters. And of course, you can add more by just duplicating those groups for handling each object.The X, Y, Width, and Height for each object are handling for the relocation of the object pasted on the new background. I suggest you test with a small frame number before you run full length of video. Because each object position are different, you have it relocate itSince it is an object pasted on t...
Fluxgym - A Simple WebUI Flux 1 Dev LoRA Training For Local Low VRAM
Tutorial Video : https://youtu.be/1DyfPhIkX78In this detailed YouTube video, we delve into the world of Flux Gym, a simplified Lora training tool designed to make AI training more accessible. From the creator of Pinokio AI, Flux Gym offers a user-friendly web UI for running Lora training locally on your machine with minimal VRAM support. The video walkthrough covers the front-end utilizing the Gradio UI from the AI toolkit, setting VRAM allocation, dataset preparation, and the backend logic using the Kohya script.AI Toolkit Tutorial Guide : htt...
FLUX 1 Dev ControlNet Union Pro Tutorial And Workflow
Video : https://youtu.be/0fw3Dqj0Fb0In this video, we unpack the complexities of ControlNet conditioning and showcase how the Union Pro model offers a comprehensive solution for executing various ControlNet preprocessors seamlessly. Learn how to optimize your workflow by utilizing different ControlNet modes such as Canny, tile, depth, blur, pose, gray, and low quality within a single ControlNet model file. By following our step-by-step tutorial, you'll gain a deeper understanding of how Union Pro empowers users to experiment with different ...
MiniMax - A New DiT AI Video Model Is Insanely Good!
Video : https://youtu.be/3ZGi4vp1NHMHailuo AI - 海螺 (Hǎiluó) AI released New DiT AI Video Model From MiniMax.Get ready to immerse yourself in the world of MiniMax, the Shanghai-based tech startup that is making waves with its groundbreaking AI models. MiniMax has recently unveiled a range of impressive AI technologies, including API-based models for voice, text-to-speech, big language models, music models, and now, an AI video generator. The best part? This AI video generator is accessible to all end users through a user-friendly web inte...
How To Run Qwen 2 VL In ComfyUI? The Best Vision Language Model Of 2024?
How To Run Qwen2 VL In ComfyUI? We are going to test out Qwen2 VL 7B in ComfyUI locally.Tutorial Video : https://youtu.be/lt_zFQY9CxkQwen2-VL Analysis : https://thefuturethinker.org/qwen2-vl-the-best-vision-language-model-of-2024/In this video, we explore the remarkable capabilities of Qwen2-VL, developed by the innovative team at Alibaba Cloud. From state-of-the-art image processing to long-form video comprehension and agent-like functionalities, Qwen2-VL sets a new standard in vision-language AI technology. Join us as we delve into the advanc...
Quick Tips - Flux 1 Dev using GGUF Quantization
Hello everyone!So ComfyUI have a new custom node recently.It allow ComfyUI (latest version) to support GGUF Quantization AI model.Some Info. about GGUF Quantization :https://thefuturethinker.org/what-is-gguf-quantization-why-it-is-fast-and-memory-efficient-inference/ComfyUI-GGUF : https://github.com/city96/ComfyUI-GGUFGGUF format popularized by llama.cpp.And using a diffusion model like Flux , it helps to reduce the VRAM usage in the execution.This is one of the image generation processes.This only things you have to change on Flux workflow, Un...
ComfyUI With Mimic Motion Change Video Background
Video Tutorial : https://youtu.be/kBUdLRWFmI8Things you need to know:MimicMotion Blog : https://tencent.github.io/MimicMotion/Github : https://github.com/Tencent/MimicMotionMimicMotion ComfyUI Custom Node: https://github.com/kijai/ComfyUI-MimicMotionWrapperWorkflow for Patreon Supporters : https://www.patreon.com/posts/111263092/
ComfyUI With Mimic Motion Change Video Background For Motion Scenes (For Patreon Supporters Only)
Video Tutorial : https://youtu.be/kBUdLRWFmI8Other Related Workflows and Tutorials:Stable Diff Mimic Motion Live Portrait Workflow V3 - For Patreonhttps://www.patreon.com/posts/stable-diff-live-109032071MimicMotion_workflow_img2img-RestyleFrame-v1 (For Patreon Supporters)https://www.patreon.com/posts/mimicmotion-v1-107525816Things you need to know:MimicMotion Blog : https://tencent.github.io/MimicMotion/Github : https://github.com/Tencent/MimicMotionMimicMotion ComfyUI Custom Node: https://github.com/kijai/ComfyUI-MimicMotionWrapperJoin us as w...
Just For FUN :) Mimic Deadpool
Actually, doing a character with mask on is pretty chill. We don't have to take care of the face 😂 , not refiner, not faceswap need.P.s : by the way, new video and workflow update tomorrow for this. It's not going to be only for dance video. Saw people asked, so with the new version workflow, I will do a short motion scenes with Mimic Motion, AnimateDiff.
HeadGAP , KEEP , Sketch2Scene - 3 New AI Project Worth Following In 2024
https://youtu.be/FR8WbwZaL2wIn this exciting video, we delve into the latest AI breakthrough projects recently unveiled on Github. From revolutionary 3D avatar creation with HeadGAP to advanced video face restoration with KEEP, and the magical transformation of sketches into 3D environments with Sketch2Scene - we explore the cutting-edge technology shaping the future of digital innovation. Join us as we break down the complexities of these AI models, discuss their real-world applications, and analyze the potential impact on various industries. ...
Let's try single character first, then multiple
BG change , static or video background. I don't like static image background, so I decide to use video.Have not make those edge on the character look nice, just a draft of it and see how a workflow could work
CogVideoX 5B With AnimateDiff V2V Upscaler Workflow (Patreon Supporters Only)
Related Post : https://www.patreon.com/posts/111067576Video : https://youtu.be/v9KGQoaqhkwHere's some step for install ComfyUI-CogVideoXWrappergit clone https://github.com/kijai/ComfyUI-CogVideoXWrapperIn CMD , locate in ComfyUI main folder location :For Comfy Portable:python_embeded\python.exe -m pip install -r ComfyUI\custom_nodes\ComfyUI-CogVideoXWrapper\requirements.txtpython_embeded\python.exe -m pip install --pre onediff onediffxpython_embeded\python.exe -m pip install nexfortFor Comfy Non-Portable:pip install -r ComfyUI\custom_nodes\...
CogVideoX 5B In ComfyUI Better Quality - Local AI Video Model That Truly Works!
Video : https://youtu.be/v9KGQoaqhkwIn this engaging video, we delve into the latest advancements in AI technology with the CogVideoX models, specifically exploring the transition from the 2B to the improved 5B version. Witness firsthand the enhanced quality, smoother animations, and more coherent styles in the 6-second video clips generated by the updated CogVideoX 5B. Follow along as we demonstrate how to implement and run these models using ComfyUI, providing insights into the installation process and performance enhancements.Customized Work...
Sneak and peek - CogVideo 5B improvement
They do listen to the market. After 2B review, they message me in X and said this 5B new release won't disappointed.Tomorrow another video about this model , and we build a local Kling, Runway, Sora, whatever you want that call.
Multi-characters & BG Vid/Img Change (Mimic) - In Progress
what do y'll think :)In Progress... hopefully can caught up this weekupdated:
One of my Video Ad Competition Entry
Here's My 2024 AI Advertising Competition Entry Video Ad, the event co-hosted by Curious Refuge and Vica.If you like it, please give this video a Vote >>> https://woobox.com/7e25kg/gallery/kZX3ZWDGoY0I only have limited time to do this one (only a day , because I need to work on other stuffs and tutorial videos.)But it was fun while playing and enjoying video creation.P.S: this is for a brand that is doing water, they are using reusable water bottle as selling point. As I like nature and less plastic, I decided to work the ad for this...
ControlNeXt SVD & Mimic Motion Comparison Workflows - For Patreon Supporters
Check out the tutorial video for the detail how it run these 2 workflows.https://youtu.be/gNY_gJ9ekyUPublic Post For All Members: https://www.patreon.com/posts/110821109/Related posts:Mimic Motion Tutorial And Workflows:ComfyUI Create AI Influencer Dance Video With Mimic Motionhttps://www.patreon.com/posts/comfyui-create-107532705MimicMotion_workflow_img2img-RestyleFrame-v1 (For Patreon Supporters)https://www.patreon.com/posts/mimicmotion-v1-107525816Stable Diff Mimic Motion Live Portrait Workflow V3 - (For Patreon Supporters)https://www.patreo...
ControlNeXt SVD An Animate AI Framework Better Than Mimic Motion? Tutorial & Comparison
https://youtu.be/gNY_gJ9ekyUControlNeXtSVD An Animate AI Framework Better Than Mimic Motion?Additional Workflow Contents for Patreon Supporters : https://www.patreon.com/posts/110821492Related posts:Mimic Motion Tutorial And Workflows:ComfyUI Create AI Influencer Dance Video With Mimic Motionhttps://www.patreon.com/posts/comfyui-create-107532705MimicMotion_workflow_img2img-RestyleFrame-v1 (For Patreon Supporters)https://www.patreon.com/posts/mimicmotion-v1-107525816Stable Diff Mimic Motion Live Portrait Workflow V3 - (For Patreon Supporters)htt...
CogVideoX - Local AI Video Generator - Too Good To Be True?
Video : https://youtu.be/ChS4JGijwPkIn this video where we delve into the AI-generated video models, CogVideoX. Join me as we explore the key features, implications, and examples of this cutting-edge technology, offering insights into its capabilities and limitations.In this video, we break down the development of CogVideoX by researchers from Zhipu AI and Tsinghua University, highlighting its advancements in video length, coherence, and motion complexity. We'll also discuss the hardware requirements for running CogVideoX locally and provid...
Updated ComfyUI UI, Menu Bar, Setting Panel
So latest ComfyUI version.You can try out with the new update UI in ComfyUI.The old floating menu on the right side is now on the top or bottom.You can choose to use this or the older version menu.Of course , you can always disable the new UI menu bar and use the old one.There are new Node Library on the side, which you can browse any nodes that you have installed. The Double Click Popup Search Bar with nodes preview on the side.The Save Workflow panel looks like ComfyUI comes with file management for Workflows.But actually, it is just going ba...
Flux 1 Dev IPAdapter Just Drop Today - Installation Tutorial Guide
https://youtu.be/JYNKx1rLpTYThenew Flux 1 Dev IPAdapter Just Drop Today.Here is a new video about it. (Don't feel hype about it yet, it's just a beta)https://huggingface.co/XLabs-AI/flux-ip-adapter
Flux 1 Dev LoRA Training Locally Using AI-ToolKit - Tutorial Guide
Tutorial Guide : https://youtu.be/F-7gfqSP2ZYIn today's video, we dive into the fascinating world of Flux diffusion models and training LoRA models to create stunning images. Join me as we explore the AI Toolkit GitHub project and learn how to install and configure the necessary components for this cutting-edge technology. From setting up the environment to training the models, we cover it all in detail. Stay tuned for a deep dive into the world of AI video scenes and image generation with Flux!Additionial Tutorial Related To Flux LoRA Trai...
Flux 1 Dev ControlNet V3 Just Drop! Installation Tutorial Guide
Tutorial Video: https://youtu.be/1t-nvX8FEqUFlux 1 Dev ControlNet V3 Just Drop! This is a tutorial of how to install it into ComfyUI and run with Flux 1 Dev.In this exciting video, we're diving deep into the world of XLab AI with the latest release of Flux One developer models. Join me as we explore the newly launched ControlNet versions for canny, soft edge, and depth map in version three. We'll walk through the installation process using the user-friendly ComfyUI manager, making it a breeze to connect custom nodes to the Flux models.G...
Experimenting ControlNeXt-SVD
ControlNeXt-SVD is very similar to Mimic Motion. They both use DWPose to create animate character motion, using SVD as animate model for the motion.But ControlNeXt-SVD , I feel like getting lower quality from it.So I am trying to use the good old AnimateDiff to refine the video after 1st pass from ControlNeXt-SVD.New video about this will post soon.
Falcon-Mamba 7B State Space Model Beat Llama 3.1 And All Same Size LLM?
I have to share about this LLM, because I am a believer of Mamba architecture for AI model.Video: https://youtu.be/4yK7tbO4keUMamba State Space Language Model : https://github.com/state-spaces/mambaFalconMamba Blog Post : https://huggingface.co/blog/falconmambaOfficial Site: https://www.tii.ae/FalconMamba Hugging Face : https://huggingface.co/tiiuae/falcon-mamba-7bFalconMamba Playground : https://huggingface.co/spaces/tiiuae/falcon-mamba-playground
Friendly Reminder - Beware of spam link
Sometime it happen in our Discord server, new joint account within a day and send out of topic links. I set an AI to auto ban those now.Just want to friendly remind you guys in Discord or other communities. Be careful with anonymous links. I know most of you are good people, and want to learn something about AI everyday.And spammer/ scammers will take this opportunity.This is one of the example:Stay safe on the Internet :)
