YouTube : https://www.youtube.com/SECourses
Patreon : https://www.patreon.com/SECourses
Please join discord and message me there so I can give you Patreon rank in our channel
Discord : https://discord.com/servers/software-engineering-courses-secourses-772774097734074388
Your Patreon support is tremendously important for me so thank you so much
https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Patreon-Posts-Index.md
https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Amazing-Generative-AI-Scripts.md
https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Generative-AI-Updates-And-News.md
- I have pre-compiled diffoctreerast and diff_gaussian_rasterization for Python 3.10
- When you download attached zip file from below post check out
- Windows_Install.bat for Python 3.10
- Just use any word editor to see their content
- Zip file : https://www.patreon.com/posts/117470976
- This zip file contains Python venv installers for TRELLIS from Microsoft to generate 3D assets from Images as GLB file and also Gaussian Splatting ply
- I have pre-compiled insightface both for Python 3.10 and 3.11 for Windows systems with CUDA 12.4+
- When you download attached zip file from below post check out
- Windows_Install.bat for Python 3.10 and
- Windows_Install_Python_3_11.bat for Python 3.11
- Just use any word editor to see their content
- Zip file : https://www.patreon.com/posts/105023709
- This zip file contains Python venv installers for ComfyUI, FaceID, IP-Adapter, InsightFace Automatic Installers for Windows, RunPod, Massed Compute, Linux
- We are using pre-compiled MultiScaleDeformableAttention wheel for Live Portrait installer
- Download zip file and edit Windows_Start_app.bat file and you will see how to use
- The zip file shared in following link : https://www.patreon.com/posts/119254105
- Let's say you are starting a batch multi-process processing in a Python app
- Having synchronized data transfer and statistics between processes is too hard
- This below Joy_Caption_Alpha_Two.py starts multiple batch captioning on each given GPU and has necessary code and custom class Statistics to ensure proper data synchronization between multiple started processes
- https://www.patreon.com/posts/118827960
- Tested on Windows and works great - should work perfectly on Linux as well
- If you are having problem with ONNX GPU Runtime due to cuDNN mismatch here the solution
- The RunPod installer script (RunPod_Install.sh) in below post has accurate install instructions and commands to install cuDNN and set as default
- https://www.patreon.com/posts/119081500
- Tested on Roop Unleashed on RunPod
- If you are having errors related torio package this is the scripts you need
- Script link : https://www.patreon.com/posts/117990364
- This script has automatic downloaders and accurate folder installers for missing FFmpeg DLL files
- This script solves the following errors
-
- ImportError: Failed to intialize FFmpeg extension. Tried versions: ['6', '5', '4', ''].
-
- File "venv\lib\site-packages\torio_extension\utils.py", line 122, in _find_ffmpeg_extension
- With Cloudflared, for free, you can access any app that is running on a remote PC via your browser in your PC, Phone, Tablet, TV etc
- Look at the Massed_Compute_Instructions_READ.txt and Massed_Compute_Install.sh to learn how to use Cloudflared amazing tunneling feature
- Way better than ngrok and it is for free no bandwith limit
- Patreon post : https://www.patreon.com/posts/117326651
- Normally downloads from Hugging Face repos are limited to 40 MB per second
- With the strategy used in download models script in the below post attached zip file, you can download model from Hugging Face Ultra fast without any speed limitation
- Kohya FLUX Fine Tuning (Full Checkpoints) Training Full Tutorial For Local Windows and Cloud RunPod and Massed Compute
- Patreon post : https://www.patreon.com/posts/112099700
- The attached Kaggle notebook is using this special strategy to download models uber fast on Kaggle
- Would work on every platform
- Free Kaggle Account Notebook for SwarmUI with FLUX, SD 1.5, SDXL & Stable Diffusion 3.5 Large and Dual T4 GPU support
- Patreon post : https://www.patreon.com/posts/106650931
How to Install CUDA 12.4 and set as default - easy - also installs cuDNN latest and works with LivePortrait XPose/models/UniPose local building
- Massed_Compute_Install.sh file has it with sudo
- Patreon post : https://www.patreon.com/posts/107609670
- Windows_Fix_cuDNN_Errors_Run_As_Administrator.bat does that so it is an example code
- https://www.patreon.com/posts/103765029
- This method can be used to install any cuDNN files directly into venv
- Tested on famous FaceFusion 3 - works perfect
- The code is inside Windows_Install.bat file
- https://www.patreon.com/posts/103765029
- The included zip file has Massed Compute instructions that explains step by step
- The instructions can be used for any Ubuntu Machine with IP
- https://www.patreon.com/posts/112912425
- Both Windows and Ubuntu 1-click installation
- Fully batch processing
- 79 Classes of Yolo V10 available
- Auto zooming in target and cropping target without cropping any part of the target
- Full automatic focus face auto resizing
- Ultimate Batch Image Pre-Processing Gradio APP for Windows and Cloud (RunPod , Massed Compute) With Yolo Models - 1-Click To Install And Use
- https://www.patreon.com/posts/112126955
- Massed_Compute_Install.sh file has it with sudo
- Runpod_Install.sh file has it without sudo
- Patreon post : https://www.patreon.com/posts/103765029
- Download zip file for FaceFusion NEXT they are inside it
Solution For Error : OSError: WinError 126 The specified module could not be found. Error loading "venv\lib\site-packages\torch\lib\fbgemm.dll" or one of its dependencies
- The solution for this error posted on below link
- https://www.patreon.com/posts/109913645
- Check the section Windows Requirements
- Hugging Face Upload / Download Notebook - Supports Private Repositories and Multi Commit as Well
- https://www.patreon.com/posts/104672510
Utra Fast and Robust and With Resume Capability Hugging Face Repo Downloader With snapshot_download Script
- Most Advanced 1-Click DeepFake / FaceSwap App Rope Installer for Windows and Massed Compute and Linux
- Rope Live has it
- https://www.patreon.com/posts/105123768
- JoyCaption : Amazing Image Captioner - Batch Processing Included - Windows, RunPod, Massed Compute 1-Click Install
- https://www.patreon.com/posts/joycaption-image-110613301
- JoyCaption : Amazing Image Captioner - Batch Processing Included - Windows, RunPod, Massed Compute 1-Click Install
- https://www.patreon.com/posts/joycaption-image-110613301
- FaceFusion (Most Advanced 1-Click DeepFake) 1-Click Installer for Windows, RunPod and Massed Compute
- https://www.patreon.com/posts/103765029
- ResShift 1-Click Windows, RunPod, Massed Compute, Kaggle Installers with Amazing Gradio APP and Batch Image Processing
- URL : https://www.patreon.com/posts/110331752
- These ones can be used as a base template
- ResShift 1-Click Windows, RunPod, Massed Compute, Kaggle Installers with Amazing Gradio APP and Batch Image Processing
- URL : https://www.patreon.com/posts/110331752
- In Massed_Compute_Install.sh file there is necessary code to install missing cuDNN CUDA dll libraries that will solve ONNX library errors when using FaceFusion and such apps
- URL : https://www.patreon.com/posts/103765029
- If you ever need to compare different files in given 2 folders with also listing missing or extra files this script is amazing
- I used Compare_2_Folders_Differences.py file inside LivePortrait installer zip to compare 2 different forks of LivePortrait repos
- URL : https://www.patreon.com/posts/107609670
- For AI repos compiling accurate gcc and g++ is a nightmate
- LivePortrait Massed Compute installer has it for XPose UniPose ops library compiling
- URL : https://www.patreon.com/posts/107609670
- On LivePortrait app, we have HF_model_downloader script which downloads given Hugging Face repo files 1 by 1 in a very robust way
- It has skip fully downloaded, retry and resume capability
- URL : https://www.patreon.com/posts/107609670