InstantID Gradio App: 1-Click Face Transfer With Xinsir ControlNet (Patreon)
Downloads
Content
InstantID Gradio app transfers a face from one photo into new SDXL images with Xinsir ControlNet; block swap lowers 1280x1280 runs to 8.25 GB VRAM.
Ultra advanced app with automatic installers for Windows, RunPod and Massed Compute and a free Kaggle notebook, now with RTX 5000 support. Uses the very best Xinsir Canny, Depth and OpenPose ControlNet models, style and model loops, LoRA support and head-only control.
Patreon exclusive posts index to find our scripts easily, Patreon scripts updates history to see which updates arrived to which scripts and amazing Patreon special generative scripts list that you can use in any of your task.
Join discord to get help, chat, discuss and also tell me your discord username to get your special rank : SECourses Discord
Please also Star, Watch and Fork our Stable Diffusion & Generative AI GitHub repository and join our Reddit subreddit and follow me on LinkedIn (my real profile)
=======
Latest zip file : instant_id_v27.zip
14 July 2025 Update
The app now supports RTX 5000 series on Windows, RunPod and Massed Compute
A lot of libraries updated and I had to do a lot of fixing for this
I have implemented block swapping feature successfully and this is a great milestone since I plan to implement it into our other apps hopefully such as upcoming video upscaler STAR model
New VRAM usages as below for 1280x1280px
Options VRAM usages are as below. No controlnet or LoRA used so these are min requirements
1+1 = 14150 MB (block swap 8 = 11000 MB)
1+3 = 9550 MB (block swap 8 = 8250 MB)
2+1 = 13500 MB (block swap not works)
2+3 = 9950 MB (block swap not works)
If you have 8 GB GPU still may work so give it a try
Models reuploaded into a HF_XET repo for faster and more robust download
Now we auto download RealVisXL V5 instead of V4
Make a fresh install for this version
10 January 2025 Update
New downloader bat file Optional_Download_Older_Control_Net.bat added
This will download original older Canny models instead of newest Xinsir
It will overwrite downloaded Xinsir models - Xinsir ControlNet models downloaded during install
I personally prefer Xinsir since they are the best ControlNet models for SDXL
Windows Requirements
Python 3.10, FFmpeg, CUDA 12.8, cuDNN 9.7, C++ Tools, MSVC and Git
If it doesn't work make sure to below NEW tutorial and install everything exactly as shown in this below tutorial
Unique Features We Have
I have made huge updates with Version 24
Now we are using Xinsir ControlNet models (Canny, Depth and OpenPose) which are the very best Controlnet models for SDXL models
I have added a new model downloader which downloads models extremely fast like 1000MB + on cloud services - you can see same speed on your Windows PC too if you have sufficient internet speed
This Gradio app is developed by me with so many additional features no other apps has
Image Metadata and Model Downloader tabs are fully working
It has so many styles and Generate With All Styles (Loop) and Generate With All Models (Loop)
You can even test and compare different LoRA + Model combinations
Change applied LoRA Scale is also fully working
How many Images to Generate fully working
One of the very best feature we have is applying certain Controlnet to the only head of the input image to increase resemblance and have better output image quality that no other apps have (Apply Head-Only Control to)
The fast model downloader will download by default below models - both are best for realism
RealVisXL_V5.0.safetensors
eldritchPhotography_v1.safetensors
Windows, RunPod and Massed Compute installers are all updated to latest Torch, xFormers, bitsandbytes, transformers, triton and ulra fast model downloaders
Kaggle notebook also completely updated and made extremely faster to download and initialize
Installers are seamlessly working on Windows, Massed Compute (recommended cloud), RunPod and on a free Kaggle account perfectly
It installs and downloads all models automatically
None of the current zero-shot face transfer into a new scene models are working as good as InstantID + xinsir SDXL ControlNet models when you need realism - stylization is another topic
RunPod Instructions
Please register for RunPod at : https://get.runpod.io/955rkuppqv4h
Select a GPU with a minimum of 24 GB VRAM, such as the RTX 3090.
Get a machine with minimum 48 GB RAM, getting multiple GPU machine better you will get more RAM
Choose a RunPod template:
RunPod Pytorch 2.2.0
runpod/pytorch:2.2.0-py3.10-cuda12.1.1-devel-ubuntu22.04
Follow Runpod_Instructions_READ.txt
Massed Compute Instructions
If you get 2x A6000 GPU it would work best with 96 GB RAM but not mandatory
Watch this instructional video to learn how to use Massed Compute:
Register for Massed Compute using the following link:
Important Notes:
Ensure each copy-paste action is accurate, as Massed Compute may not always copy-paste correctly.
Select an A6000 GPU or A6000 GPU ALT (if other one is not available) and use our code SECourses for huge discount.
Follow Massed_Compute_Instructions_READ.txt
If you want to see history of development please read this obsolete (not updated and used anymore) topic : https://www.patreon.com/posts/1-click-very-gui-97769887
Related posts
FaceFusion 3.5 1-Click Installers: DeepFake App for Windows & Cloud
VisoMaster Fusion 1-Click Installer: 0-Shot Face Swap & Deep Fake App
IP-Adapter-FaceID-PlusV2: 0-Shot Face Transfer Auto Installer & App
All posts in this topic: Face Swap, Deepfake and Talking Avatar Apps · New to SECourses? Start Here: Best and Most Popular SECourses Posts