Skip to content

Turn Videos Into Animation 3D Just 1 Click ReRender A Video Tutorial Installer For RunPod

Jump to bottom
FurkanGozukara edited this page Oct 23, 2025 · 1 revision

Turn Videos Into Animation / 3D Just 1 Click - ReRender A Video Tutorial - Installer For RunPod

Full tutorial link > https://www.youtube.com/watch?v=cVf9Qf_pKks

Turn Videos Into Animation / 3D Just 1 Click - ReRender A Video Tutorial - Installer For RunPod

image Hits Patreon BuyMeACoffee Furkan Gözükara Medium Codio Furkan Gözükara Medium

YouTube Channel Furkan Gözükara LinkedIn Udemy Twitter Follow Furkan Gözükara

In this video I share automatic RunPod installer for ReRender A Video github repository - open source AI project. The auto installer script is here : https://www.patreon.com/posts/1-click-auto-for-91039997

Auto script is here ⤵️

https://www.patreon.com/posts/1-click-auto-for-91039997

Official repo is here ⤵️

https://github.com/williamyang1991/Rerender_A_Video

SECourses Discord To Get Full Support ⤵️

https://discord.com/servers/software-engineering-courses-secourses-772774097734074388

My LinkedIn ⤵️

https://www.linkedin.com/in/furkangozukara/

My Instagram ⤵️

https://www.instagram.com/gozukarafurkan/

My Medium ⤵️

@FurkanGozukara https://medium.com/@furkangozukara

My CivitAI ⤵️

https://civitai.com/user/SECourses

Video Transcription

  • 00:00:00 Greetings everyone. In this video I am going to  show you how to use ReRender A Video and new open  

  • 00:00:08 source repository to generate amazing animation  videos, 3d videos by using existing real videos.  

  • 00:00:16 If you don't know what is it let me show you. For  example I have converted this Larry Wheels clip  

  • 00:00:23 into this animation clip with just one click.  So I will show you how to turn your real videos  

  • 00:00:31 into animation or 3d on RunPod with one click  installer. All you need to do is download this  

  • 00:00:37 installer.sh file open your RunPod from this link.  If you don't have a RunPod sign up or login. If  

  • 00:00:44 you have let's login. Then go to community cloud,  select extreme speed from here. You can use higher  

  • 00:00:52 VRAM having machines with higher batch size.  However, for this tutorial I will use RTX 3090  

  • 00:01:00 which is one of my best GPU selection. From here  type fast and select RunPod Fast Stable Diffusion  

  • 00:01:07 template. Click continue. Deploy. Go to my pods  wait until connect button appears here. Okay  

  • 00:01:15 connect button appeared, click connect. Connect  to the Jupyterlab. Once the interface has opened  

  • 00:01:20 click this icon, upload the installer.sh file  then all you need to do is start a new terminal  

  • 00:01:27 by clicking here and use these two commands. So I  copy them, copy paste them and hit enter. And it  

  • 00:01:36 will start the fully automatic installation for  you. It will also download the necessary model  

  • 00:01:42 files automatically for you. So the necessary  libraries installation have been completed.  

  • 00:01:47 Now it is downloading the models automatically  for you into the correct folders. All right,  

  • 00:01:53 all of the models are also downloaded. Now we  are ready to launch the web UI. To do that,  

  • 00:01:59 we will use this part of the script. So let's open  a new terminal here. Copy paste everything. And  

  • 00:02:07 let's say if you want to use additional models,  what you need to do is: you need to download  

  • 00:02:14 them inside models folder. And you need to modify  sd_model_cfg.py file and add them here. As you are  

  • 00:02:23 seeing right now. Moreover, I suggest you to edit  this video_blend.py file and make OPEN_EBSYNTH_LOG  

  • 00:02:33 True. This is super important because with  this you will have much more verbose output  

  • 00:02:40 in the workspace where you have launched. Now it  has started, we will use Gradio Live interface.  

  • 00:02:47 Currently, we are not able to upload our video  into workspace and give its path. I have opened  

  • 00:02:54 an issue thread for this. So we are going to use  upload from here. Let's select our video. So let's  

  • 00:03:01 go to video file location. I will use this video.  Where was it? Here. So this is the video. First of  

  • 00:03:10 all, you need to define your scene very carefully.  If your scene has multiple different scenarios,  

  • 00:03:19 then you should split them because your prompt  should be able to define your scene as much  

  • 00:03:24 as possible. In this scene, a bodybuilder is  posing in a gym, so a muscular man is posing  

  • 00:03:33 in a gym. Then there are several settings. Default  settings are really good. But if your video is HD,  

  • 00:03:41 I suggest you to make this 704. Then I suggest  you to change this to Canny. Make Canny low  

  • 00:03:48 threshold 25, high threshold 50. Moreover,  your base model will have huge effect. So you,  

  • 00:03:55 I suggest you to use Flat_2D_Animerge, you  can also try different models. This is an  

  • 00:04:01 anime model. And most importantly, according to  your resolution, you need to reduce this number  

  • 00:04:07 of parallel processes. For 24 gigabyte and this  resolution 2 is the maximum that you should set.  

  • 00:04:15 If you go higher resolution, make this 1, if you  go lower resolution, and according to your VRAM,  

  • 00:04:20 you can increase this. This will significantly  affect your processing time. So it is super  

  • 00:04:26 important. And before clicking run all, make sure  that your first keyframe is what you are looking  

  • 00:04:33 for. When you first time hit the first keyframe,  it will load all of the models as you are seeing  

  • 00:04:39 right now. We are able to watch everything in this  started terminal. The first time will be slower,  

  • 00:04:45 but second and third and more will be faster.  Because it will cache it in the memory. By the  

  • 00:04:52 way, since this is running on a remote server, it  will upload file to the remote server. And if your  

  • 00:04:59 file is too big, it will take a lot of time. Once  your file is uploaded, these options will become  

  • 00:05:05 available. The number of key frames. I have  tested different key frames, and 10 is really,  

  • 00:05:10 really optimal. So you will see a loading icon  here. I will also demonstrate it. Okay, we got the  

  • 00:05:16 first key as you are seeing right now. For example  we can also try with another model. Let's try with  

  • 00:05:23 RevAnimated. It is also downloaded and let's  say a muscular man is posing in a gym 3d CGI,  

  • 00:05:30 okay, character. Let's see the first key frame  output for this. So it will reload the different  

  • 00:05:37 model. And we got the output. It is really really  looking like a 3d CGI character as you are seeing  

  • 00:05:45 right now. Actually it is amazing. It is even  better looking than the other one. Let me show  

  • 00:05:51 you what I mean by uploading, uploading process.  For example I have another file. Let's see, here:  

  • 00:05:59 this is three megabytes and this is 12 megabytes.  Okay, I selected it. And after I selected it you  

  • 00:06:06 see this part of the web UI became like this.  It is loading. So this means that it is actually  

  • 00:06:13 uploading. You need to wait here become available.  Once it becomes available It means that the file  

  • 00:06:20 is uploaded. I can also see the upload process  here. And it is uploaded. Very nice. So it was  

  • 00:06:26 fast. Now I can click run first keyframe and it  will make this a first keyframe. Let's see also  

  • 00:06:33 that one. And we got the first keyframe so we can  run all and process it all. So it is also looking  

  • 00:06:39 like a CGI character. And let's return back to  our previous one. And now it will re-upload.  

  • 00:06:47 Every time it re-uploads unfortunately. I opened  an issue for this. Maybe it will get fixed in the  

  • 00:06:53 future versions. So where will be the output  file? The output video will be here once it  

  • 00:06:59 is completed. But let's say the Gradio app lost  its connection. So you will find the file inside  

  • 00:07:06 render a video folder. Inside results: you will  see the file name, the video file name having  

  • 00:07:13 folder here. So it will be inside there. For  example, the for the posing video, we will have  

  • 00:07:19 a blend.mp4 file here. Okay, let's see, the  file is uploaded. Let's click first keyframe.  

  • 00:07:30 Okay. Now I am ready to process all. Before doing  that I will make sure that the number of parallel  

  • 00:07:36 processes are 2 and run all. And that's it. Now  I need to wait it to process all of the keyframes  

  • 00:07:44 and everything and the result will be inside this  folder. We can also see the keyframes here in a  

  • 00:07:50 moment. Now it generated the 11th frame. Yes. Now  we can see it. Okay, it started generating the  

  • 00:07:58 keys. So it is working. And as I said, the final  video will be here. Then we can download. I also  

  • 00:08:05 have added how to merge the output video with  its sound. Unfortunately, it is not keeping the  

  • 00:08:12 sound. Moreover, if your video FPS is like this,  you may want to convert it to closest FPS with  

  • 00:08:21 FFMPEG. Thank you very much for supporting me on  Patreon. Hopefully I will make a public tutorial,  

  • 00:08:26 step-by-step tutorial for my non-Patreon  supporter subscribers as well. See you later.

Clone this wiki locally