Descriptions:
Bernini-R now has native support in ComfyUI and provides a great way for you to do both image and video editing.
Bernini-R uses the Wan2.2 architecture, employing two separate transformer models for high and low noise denoising, and so the existing lightx2v LoRAs work well to speed things up for any task. It utilizes SA3D RoPE segment-aware 3D rotary position embeddings to manage temporal consistency, and supports multiple references images – or even both a source and reference video combination.
In this tutorial you get a hands-on human review demonstrating how the reference-guided frame-by-frame image swap actually handles complex objects in a real ComfyUI workflow.
You’ll learn about various nodes to both improve image quality, and reduce VRAM requirements – great if you have low VRAM. ComfyUI have provided both fp8 and fp16 model versions, with the fp8 model quality being just fine with half the system requirements. Plus, if you’re editing images but want a higher resolution output, then see the WAN 2x upscaler VAE in action to produce images at 2560×2560!
Edit images, do video face swaps, insert or remove objects in videos, and much, much more – all in single workflow that you can build for yourself and run on your own home computer 🙂
== Get Workflows! ==
Looking to support the channel? Or perhaps you want to have the featured workflow READY-MADE for you? Skip the D.I.Y & jump straight in using the link below:
https://www.patreon.com/NerdyRodent/posts/precise-video-161442856
== Join this channel to get access to perks ==
https://www.youtube.com/channel/UC4-5v-f-xKnbi1yaAuRSi_w/join
== More Links ==
ComfyUI – https://github.com/Comfy-Org/ComfyUI
Bernini GitHub: https://github.com/bytedance/Bernini
Bernini Project Page: https://bernini-ai.github.io
Original Bernini-R Model Weights: https://huggingface.co/ByteDance/Bernini-R
ComfyUI Bernini-R Model Weights: https://huggingface.co/Comfy-Org/Bernini-R
Bernini Paper (arXiv): https://arxiv.org/abs/2605.22344
Wan 2.2: https://huggingface.co/Wan-AI
Lightx2v LoRAs: https://huggingface.co/lightx2v
== Comfy Cloud ==
Worried your computer isn’t up to the job? Think you don’t have enough RAM or VRAM and don’t want to even try on your own PC? Then why not jump straight to Comfy Cloud – the easiest way to use Comfy online without worry!
https://www.comfy.org/cloud
(Now with FREE tier! – https://blog.comfy.org/p/free-tier-arrives-in-comfy-cloud)
== Beginner’s Guides! ==
Just starting out? ComfyUI has been around for a while now, but my older videos still stand strong and provide a firm base to start from.
* LEARN how to get Comfy installed the RIGHT way.
* Avoid those weird, obfuscated installers which stunt your learning.
* Skip the confusion + any strange errors in the future from using an installer.
* UNDERSTAND the workflow creation process.
1. Installing Anaconda for MS Windows Beginners – https://youtu.be/OjOn0Q_U8cY
2. Installing ComfyUI for Beginners – https://youtu.be/2r3uM_b3zA8
3. ComfyUI Workflows for Beginners – https://youtu.be/VM9snsuoqBc
== The VERY BEST ComfyUI Playlist! ==
Is this tutorial too advanced? Do you need to learn more before you start with Comfy? Don’t worry – I’ve got YEARS of video tutorials! Go back to where it all started and catch up in no time – from a REAL HUMAN TUTOR:
== Computing ==
🔧 For a viable AI experience using your own hardware, I suggest:
* Ubuntu (I am because we are) 🐦
* An Nvidia GPU (16 GB VRAM min, 24+ GB better)
* 64+ GB RAM
* 2+ TB SSD
* CPU with 6+ cores
== Contents ==
0:00 Bernini-R ComfyUI Workflow Details
7:34 Bernini-R Image Editing and Upscaling
10:51 Bernini-R Video Editing and Face Swaps
14:15 Bernini-R Video Interpolation
