This video provides a guide for creating video clips using ComfyUI with Stable Video Diffusion, including custom workflows that improve upon existing examples, adding upscaling and a great solution for frame interpolation.
* The "Improved" workflows below include tweaks discovered after releasing this video to improve quality of outputs.
Links
Custom Workflows: drive.google.com/drive/folder...
*Improved txt2img2vid Workflow: comfyworkflows.com/workflows/...
*Improved img2vid Workflow: comfyworkflows.com/workflows/...
ComfyUI Github Repo: github.com/comfyanonymous/Com...
ComfyUI img2vid Examples: comfyanonymous.github.io/Comf...
Stability AI img2vid-xt Models: huggingface.co/stabilityai/st...
JuggernautXL Model: civitai.com/models/133005/jug...
Video Helper Suite Node: github.com/Kosinkadink/ComfyU...
How To Install Custom Nodes: stable-diffusion-art.com/comf...
OpenModelDB / LSDIRDAT Upscaler: openmodeldb.info/models/4x-LS...
comfyui_controlnet_aux(optional): github.com/Fannovel16/comfyui...
Flowframes (repo): github.com/n00mkrad/flowframes
Flowframes (itch.io): nmkd.itch.io/flowframes
Chapters
0:00 Overview
0:22 Install ComfyUI
1:02 Download Stable Video Diffusion Models
1:49 Choose a txt2img Model
2:01 ComfyUI's img2vid/txt2img2vid Examples
3:32 Custom txt2img2vid Workflow
5:26 Upscaling in the Custom Workflow
6:08 Running Through the Workflow
8:35 Custom img3vid Workflow
10:08 Frame Interpolation
12:40 Results / Comparison
*Narration created using Elevenlabs' SpeechToSpeech Synthesis
11 июл 2024