UPDATE: You can download SVD ver 1.1 here: https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt-1-1/tree/main
"This model was trained to generate 25 frames at resolution 1024x576 given a context frame of the same size, finetuned from SVD Image-to-Video [25 frames]."
From Stable Video Diffusion's Img2Video, with this ComfyUI workflow you can create an image with the desired prompt, negative prompt and checkpoint(and vae) and then a video will automatically be created with that image.
v1.3_sd3: txt2video with Stable Diffusion 3 and SVD XT 1.1
For information where download the Stable Diffusion 3 models and where put the .safetensors: https://civarchive.com/models/497255/stable-diffusion-3-sd3
v1.3_cascade: txt2video with Stable Cascade and SVD XT 1.1
For information where download the Stable Cascade models and where put the .safetensors: https://gist.github.com/comfyanonymous/0f09119a342d0dd825bb2d99d19b781c
v1.3: svd default is now: svd_xt_1_1.safetensors
.gif creation instead .webp(.mp4 is still present)
v1.2: changed the default value of fps svd img2vid_conditioning from 8 to 10, previously it was erroneously different from the save value and save webp value of 10
v1.1: more clear interface
v1: first version
Required:
1- on Windows: 8+ GB VRAM NVIDIA gpu only
on Linux: 8+ GB VRAM NVIDIA gpu and AMD gpu
2- ComfyUI: https://github.com/comfyanonymous/ComfyUI
3- this worfklow
4- in ComfyUI\models\checkpoints:
https://huggingface.co/stabilityai/stable-video-diffusion-img2vid/blob/main/svd.safetensors <-- for 14 fps video
or
https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt/blob/main/svd_xt.safetensors <-- for 25 fps video
5- a checkpoint in ComfyUI\models\checkpoints
6- a vae in ComfyUI\models\vae (optional)
7- ComfyUI Manager for the Missing Nodes: https://github.com/ltdrdata/ComfyUI-Manager
P.S.: The recommended resolution for SVD is 576x1024, but in the workflow I set a default value of 512x768 (with 40 steps, 20 steps is good enough). It's up to you to choose the resolution and steps that you want... but 512x768 with 40 steps is a good starting point for your attempts.
If you want to support me, check out my Patreon: https://www.patreon.com/Shiroppo
Follow me on civitai(!!!): https://civarchive.com/user/shiroppo
Follow Me on Twitter/X: https://x.com/ShiroppoTwit
Pixiv: https://www.pixiv.net/users/96490888
---
Description
Txt2Video with Stable Diffusion 3(Medium) and SVD
FAQ
Comments (9)
The included workflow doesn't have a save image node like shown in your example. Can you help me add this node?
Which version? I re-downloaded the sd3 version and the save image node was already inserted, anyway just copy and paste the save image node from another project and then connect it to the vae decode
@shiroppo Thanks. I was being blind and not seeing it.
Noob question, what do I need for a freshly installed ComfyUI and an SVD model. I get a VHS_VideoCombine node error, it seems like it's missing.
For comfyui, if you do not want to install it manually, you can directly download the portable version here: https://github.com/comfyanonymous/ComfyUI/releases
In the description of this workflow:
SVD 1.1(latest version): https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt-1-1/tree/main
The title is a bit misleading as it's more like txt2img2vid not txt2vid. Not a fault of the workflow of course since SVD doesn't have text guidance, and the workflow is good for what it does, just don't expect "true" txt2vid(hopefully we get a version of SVD or a similar model which will allow text guidance in the future)
is it possible to make 150 Frames ?, so that the results are similar klingai 😂😅
yeah, I'm not going to spend five hours trying to get ten dependencies to play nice to get this to work when NONE of them can be bothered to create a readme that says anything beyond 'read this marketing blurb about what our program can do' rather than what a readme is supposed be, ie 'here's how to install and setup program'
I run this workflow and while I do not get any errors I just get a black image. Any ideas?
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.