Skip to content

nerdyrodent/AVeryComfyNerd

Repository files navigation

Overview

A variety of ComfyUI related workflows and other stuff. You'll need different models and custom nodes for each different workflow. As this page has multiple headings you'll need to scroll down to see more.

Resources

You'll need models and other resources for ComfyUI. Check the table below for links to everything from ControlNet models to Upscalers

Item Description Link
ComfyUI The main thing you'll need! https://github.com/comfyanonymous/ComfyUI
See https://youtu.be/2r3uM_b3zA8 for an install guide
ComfyUI Manager Install any missing nodes using this https://github.com/ltdrdata/ComfyUI-Manager
Stability AI Models & VAEs https://huggingface.co/stabilityai
Text-to-Image models Text-2-image models https://huggingface.co/models?pipeline_tag=text-to-image&sort=trending
SSD-1B Text2-image model https://huggingface.co/segmind/SSD-1B
ControlNet Models ControlNet Models https://huggingface.co/lllyasviel/sd_control_collection/tree/main
https://huggingface.co/comfyanonymous/ControlNet-v1-1_fp16_safetensors/tree/main
QR Code Monster Control Net ControlNet Model https://huggingface.co/monster-labs/control_v1p_sd15_qrcode_monster
IP Adapter Github Repo https://github.com/tencent-ailab/IP-Adapter
IP Adapter models Models https://huggingface.co/h94/IP-Adapter
T2I Adapter Github Repo https://github.com/TencentARC/T2I-Adapter
Control LoRA Control Models https://huggingface.co/stabilityai/control-lora
AnimateDiff Original repo, many links and more info https://github.com/guoyww/AnimateDiff
Latent Consistency Models Models https://huggingface.co/latent-consistency
Upscale Wiki Many models & info https://upscale.wiki/wiki/Main_Page
Artist Style Studies SDXL Prompt output examples for inspiration https://sdxl.parrotzone.art/

List of workflows available

In ComfyUI the image IS the workflow. Simply drag or load a workflow image into ComfyUI! See the "troubleshooting" section if your local install is giving errors :)

Workflow Description Version
Basic SDXL ControlNet workflow.
Introductory SDXL Canny & Depth ControlNet example.
See https://youtu.be/reqamcrPYiM for more information.
SDXL
Basic QR Code Monster SD 1.5 ControlNet - make spiral art!
See also - https://youtu.be/D4oJz0w36ps
SD 1.5
QR Code Monster SD 1.5 ControlNet - make animated spiral art!
See also: https://youtu.be/D4oJz0w36ps
SD 1.5
Updated QR Code Monster SD 1.5 ControlNet with AnimateDiff and FreeU
Works best with the v1 QR Code Monster - https://huggingface.co/monster-labs/control_v1p_sd15_qrcode_monster
SD 1.5
AnimateDiff with Montion LoRA example. Pan up, down, left right, etc. SD 1.5
Instant LoRA 1
Inspired by AloeVera (almost identical).
Really simple, no training, "LoRA" like functionality.
SD 1.5. IP Adapter models:
1. https://huggingface.co/h94/IP-Adapter/blob/main/models/ip-adapter-plus_sd15.bin -> custom_nodes/IPAdapter-ComfyUI/models.
2. https://huggingface.co/h94/IP-Adapter/blob/main/models/image_encoder/model.safetensors -> models/clip_vision.
NB (2024). As IPAdapter-ComfyUI from 2023 is now deprecated, you should replace it with a currently supported version before use
Video guide - https://youtu.be/HtmIC6fqsMQ
SD 1.5
Instant LoRA 2
As above, but with ControlNet to guide the shape
SD 1.5
Instant LoRA 3
As above, but with QR Code Monster ControlNet too :)
SD 1.5
Instant LoRA 4
As above, but with basic upscaling
SD 1.5
Instant LoRA 5
As above, but with more upscaling to 16k+
SD 1.5
Instant LoRA 6
As above, but different upscaling to 16k+
SD 1.5
Morphing AI videos of any length using AnimateDiff. SD 1.5. Includes IPAdapter & Upscaling. IP Adapter models:
1. https://huggingface.co/h94/IP-Adapter/blob/main/models/ip-adapter-plus_sd15.bin -> custom_nodes/IPAdapter-ComfyUI/models.
2. https://huggingface.co/h94/IP-Adapter/blob/main/models/image_encoder/model.safetensors -> models/clip_vision.
Video guide - https://youtu.be/6A3a0QNPhIs
SD 1.5
Morphing AI videos of any length using AnimateDiff. SD 1.5. Includes Upscaling. Like above, but without IPAdapter controls. SD 1.5
SDXL "Instant LoRA" - basic.
Really simple, no training, "LoRA" like functionality.
Uses SDXL IP Adapter - https://huggingface.co/h94/IP-Adapter
Video - https://youtu.be/dGL02W4QatI
SDXL
SDXL "Instant LoRA" - with CLIP Vision
Uses SDXL IP Adapter - https://huggingface.co/h94/IP-Adapter
Also use "Revisions" CLIP vision - https://huggingface.co/stabilityai/control-lora
SDXL
SDXL "Instant LoRA" - with CLIP Vision & ControlNet
Uses SDXL IP Adapter - https://huggingface.co/h94/IP-Adapter
Also use "Revisions" CLIP vision - https://huggingface.co/stabilityai/control-lora
SDXL
AnimateDiff + QRCode (Vid2Vid)
Use any high-contrast input video to create guided animations! Spirals away...
SD 1.5

SD 1.5 Reposer (2 versions) - single face image to any pose. Get consistent faces!
No "roop" or similar face-swapping nodes required = easy install!
SD 1.5 ControlNet models:
https://huggingface.co/comfyanonymous/ControlNet-v1-1_fp16_safetensors/tree/main
IP Adapter models:
1. Face = https://huggingface.co/h94/IP-Adapter/blob/main/models/ip-adapter-plus-face_sd15.bin
2. Vision = https://huggingface.co/h94/IP-Adapter/blob/main/models/image_encoder/model.safetensors
NOTE Reposer2.png now uses the even more updated version of IPAdapter
Reposer Plus Bypass Edition is deprecated, but still available for download if you want to update any nodes at home.
Original Reposer Basic Video guide - https://youtu.be/SacK9tMVNUA
Original Reposer Plus Video guide - https://youtu.be/ZcCfwTkYSz8
SD 1.5
SD 1.5 Video Styler! Combining IPAdapter with Video-to-video for strange styles and weird animations
Uses https://github.com/cubiq/ComfyUI_IPAdapter_plus
The pre-trained models are available on huggingface, download and place them in the ComfyUI/custom_nodes/ComfyUI_IPAdapter_plus/models directory.
For SD1.5 you need:
* ip-adapter_sd15.bin
* ip-adapter_sd15_light.bin
* ip-adapter-plus_sd15.bin
* ip-adapter-plus-face_sd15.bin
Additionally you need the image encoder to be placed in the ComfyUI/models/clip_vision/ directory.
They are the same models used by the other IPAdapter custom nodes ;) - symlinks are your friend!
Video guide - https://youtu.be/kJp8JzA2aVU
SD 1.5
SDXL version of Reposer using the SDXL "IPAdapter Plus Face" model
Pick a face then add a body in any pose - no training!
Works with photorealistic faces, anime faces, cartoon faces, etc
SDXL
SSD-1B Workflow - SDXL for 8GB VRAM cards!
Model - https://huggingface.co/segmind/SSD-1B
Video - https://youtu.be/F-bKndyQ7L8
SSD-1B
LCM LoRA vs Normal 1.5, SDXL, SSD-1B
IPAdapter Attention Masking Example
Video https://youtu.be/riLmjBlywcg
1.5
IPAdapter Attention Masking Example with extra toppings (LCM, Facefix)
Video https://youtu.be/riLmjBlywcg
1.5
Stable Video Diffusion example with a simple upscale and frame interpolation SVD
SDXL Turbo - 1 step diffusion! SDXL Turbo, SD2 Turbo
A very basic attempt at a "Comfy MagicAnimate". Needs more work :)
Links:
Magic Animate - https://github.com/magic-research/magic-animate
Magic Animate (Windows) - https://github.com/sdbds/magic-animate-for-windows
DreamingTulpa - https://twitter.com/dreamingtulpa/status/1730876691755450572
CocktailPeanut - https://twitter.com/cocktailpeanut/status/1732052909720797524
Google Colab - https://github.com/camenduru/MagicAnimate-colab
Huggingface Space - https://huggingface.co/spaces/zcxu-eric/magicanimate
Vid2DensePose - https://github.com/Flode-Labs/vid2densepose

Model Downloads for the MagicAnimate Gradio App:
mkdir -p magic-animate/pretrained_models
cd magic-animate/pretrained_models
git lfs clone https://huggingface.co/runwayml/stable-diffusion-v1-5 -b fp16
git lfs clone https://huggingface.co/stabilityai/sd-vae-ft-mse
git lfs clone https://huggingface.co/zcxu-eric/MagicAnimate
Video - https://youtu.be/td27SyA9M80
SD 1.5
Steerable Motion - Image Batch with AnimateDiff
Video guide - https://youtu.be/bH-56e3cR2g
SD 1.5
Unsampler - Turn images into noise and back again, as modified by your prompts!
Video guide - https://youtu.be/qW1I7in1WL0
SD 1.5, SDXL
Refacer - Change the style of any face!
Video guide - https://youtu.be/r7Iz8Ps7R2s
SDXL

Troubleshooting

When troubleshooting (working to fix issues) - such as with your local custom node installs, it's best to do all of these steps until resolution.

  • Need even more help or updated workflows? Drop me a DM on https://www.patreon.com/NerdyRodent and help support the channel! :)
  • Make sure you've installed the drivers for your graphics card
  • In ComfyUI the image IS the workflow.
  • These workflows require ComfyUI to run, so you'll need to install that first. See https://youtu.be/2r3uM_b3zA8 for an install guide
  • Install ComfyUI Manager next - https://github.com/ltdrdata/ComfyUI-Manager
  • Use ComfyUI Manager to install missing custom nodes by clicking "Install Missing Custom Nodes"
  • If ComfyUI Manager can't find a node automatically, use the search feature
  • Be sure to keep ComfyUI updated regularly - including all custom nodes. Old versions may result in errors appearing. This is the most common issue, so update now!
  • These are just workflows - no custom nodes here, so no code to go wrong :)
  • Need a model or checkpoint? See the resources section!
  • By default, models are saved in subdirectories under ComfyUI/models, though some custom nodes have their own models directory.
  • Don't mix SDXL and SD1.5 models (unless stated, such as SDXL needing the SD 1.5 vision model) - chances are you'll get an error!
  • Don't try to use SDXL models in workflows not designed for SDXL - chances are they won't work!
  • Ensure your model files aren't corrupt - try a fresh download if a particular model gives errors
  • Some workflows are large. Zoom out to see more of the canvas.
  • Custom node still red after installing it? Remember to restart ComfyUI!
  • Custom node still giving an error? Check the GitHub page for that custom node - maybe someone else has a similar issue open?
  • Not sure where the GitHub page is for a custom node? You can click on it via ComfyUI Manager
  • Check the output when ComfyUI starts up as issues can show up there
  • Try updating custom nodes manually (git pull)
  • Over time, some custom nodes implement breaking changes so nodes may need to be replaced. Often this can be done with "fix node". Typically the custom node GitHub page has additional information.
  • Sometimes custom nodes just break! Check the github page for the custom node causing any issues for more information and to raise issues
  • Sometimes custom nodes change functionality, so check for updates. Changes include:
    • Segement anything - mask output inverted & now returns multiple images
    • Dynamic Thresholding - output different
    • The original IPAdapter ('IPAdapter-ComfyUI') is deprecated and has been moved to the legacy channel.
    • March 2024 - the "new" IP Adapter node (IP Adapter Plus) implemented breaking changes which require the node the be re-created. The usual "fix node" won't work.
  • The Microsoft Windows portable version of ComfyUI apparently has issues with various custom nodes, whereas normal installs are OK. Unknown error? Try a normal install!
  • Need more help? See this Playlist with loads of ComfyUI guides

Updating / Installing Custom Nodes

  • Use ComfyUI Manager to install and update custom nodes with ease!
  • Click "Install Missing Custom Nodes" to install any red nodes
  • Use the "search" feature to find any nodes
  • Be sure to keep ComfyUI updated regularly - including all custom nodes. Old versions may result in errors appearing.

Custom Node List

  • Screenshots of my installed custom nodes for reference. Not all nodes are used in workflows.
  • Install custom nodes using ComfyUI manager
  • See the Troubleshooting section if you have errors with your local ComfyUI install



About

ComfyUI related stuff and things

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published