1 Click Auto Windows Installer For Rerender A Video - 1 Click Video To Anime (Patreon)

Patreon exclusive posts index

Join discord and tell me your discord username to get a special rank : SECourses Discord

15 February 2024 Update :

• Install script updated according to latest libraries
• When installing please wait fully otherwise you will get errors
• When installing it will download a lot of models so if you don't want models to be download edit install2 .py file before starting install and delete the models you don't want
• The models it will download are : Realistic_Vision_V5.1, Flat_2D_Animerge, ReV_Animated_122 & Stable_Diffusion_1.5
• Please download latest installer all_files_V2.zip file and extract into a folder named such as c:/rerender_v2
• Do not use spaces or special characters in your folder names or in your video names 
• To see all process edit video_blend .py file and set OPEN_EBSYNTH_LOG = True
• Triton library added for Windows 
• You need to have installed Python 3.10.x - it can be like 3.10.6 or 3.10.9 or 3.10.11 etc
• Here a full tutorial for how to install FFmpeg and Python properly : https://youtu.be/-NjNy7afOQ0

19 October 2023 Update :

• Full tutorial video released > https://youtu.be/a8oeCFyM5gA
• Please also upvote this Reddit Thread Thank you so much

EXAMPLE VIDEO : https://twitter.com/GozukaraFurkan/status/1703797121126662602

Please read carefully and upvote this Reddit thread if possible

The original GitHub repo to read and learn more : https://github.com/williamyang1991/Rerender_A_Video

My installers will make a new VENV and install everything there. So this will not affect your other applications such as Automatic1111 Web UI

You also need to install ffmpeg and set its environment path. Shown in this video (2:56) : https://youtu.be/OI1LEN-SgLM

You need to have installed git

install_large_vram .bat will install Torch 2.0.1, Cuda 11.8, and xFormers 0.0.22

install_low_vram .bat will install Torch 1.13.1, Cuda 11.7 and xFormers 0.0.16 - will be probably slower i am yet to test

After install run .bat

The starting URL will be : http://127.0.0.1:7860

Before installing everything there are several crucial issues

Edit install2 .py and comment out the models that you don't want to download or you can even add new models for download. Auto download included models are Realistic_Vision_V5_1, Stable_Diffusion_1_5, Flat_2D_Animerge (which i used in example video), ReV_Animated_122

If you add new models edit sd_model_cfg .py and add them to the model dictionary list with same format. So easy to edit both of the files they are very clear

Most importantly if you don't compile ebsynth yourself please download and extract ebsynth .zip into the same folder. I compiled it with using Microsoft Visual Studio Community Edition 2022 and CUDA 12 library. Compiling it really hard. For those who wants I can connect your PC and help you if you become my Level 50 patreon subscriber (only 1 time)

Moreover, put all of these files into a folder with specific path. Do not use spaces in folder path. Do not use non english characters in folder path. Example correct folder path is

G:\render_v2

Here how your folder should look like before install

The install will take some time since it will download models which are total 32.2 GB

After all installation has been completed double click run .bat file and it will start the Web UI interface

Usage is intuitive

I made this video with default settings, flat2DAnimerge_v30 model and just a man in a room prompt

Processing is taking a lot of time

You can follow the process from CMD window

The starting URL will : http://127.0.0.1:7860

If you get ffprobe error that means it doesn't see your FFMPEG

So download ffmpeg from here : https://ffmpeg.org/download.html

Extract downloaded ffmpeg into the scripts folder (..\Rerender_A_Video\venv\Scripts) like below

Upload your video and set settings and then click Run 1st Key Frame

This will display first frame

After that click Run All Frames once you are okay with first frame

You will see the results inside Rerender_A_Video\result folder with the name of your input video file name

Make your video filename also specific easy like test_video.mp4 - dont use spaces or non-english characters

The final and most time consuming part will generate images under blend folder

It will generate number of frames equal to the number of frames in your original video

Finally it will not copy the sound of the original video

• Put output render video and initial original video into the same folder and run below command.
• Don't forget to change file names. Also you need to have installed (add system path) or downloaded ffmpeg into the same folder
• ffmpeg -i blend.mp4 -i source_video.mp4 -c:v copy -c:a copy -map 0:v:0 -map 1:a:0 output.mp4

Hopefully I will make a detailed tutorial video for this amazing tool

all_files_V2.zip contains all of the files including ebsynth .zip extracted