Consistant Multi Character AI Digital Models
The Ultimate Guide to AI Digital Model on Stable Diffusion ComfyUI course is now available! 🎊
Thank you for your patience and support. As a token of our appreciation, we're excited to offer you an exclusive 40% discount.
Simply use the coupon code "AICONOMIST40" at checkout.
This offer is valid for the next 7 days, so don’t miss out!
Don’t miss this chance to elevate your AI image generation skills!
INSTRUCTIONS
1 - Install Use Everywhere from the Manager : https://github.com/chrisgoringe/cg-use-everywhere
2 - Install Inspire Pack from the Manager : https://github.com/ltdrdata/ComfyUI-Inspire-Pack
3 - Make Sure You Downloaded All IPadapter Models and Clip Vision: https://github.com/cubiq/ComfyUI_IPAdapter_plus
4 - Restart ComfyUI
LASTEST VIDEOS:
07/08/24 - How To Install & Use LivePortrait on ComfyUI (Step-by-Step) ➡️ https://youtu.be/iMJJRKkTZo0
06/12/24 - Stable Diffusion 3 on ComfyUI: Tutorial & My Unexpected Disappointment ➡️ https://youtu.be/8goQzQMrHW0
06/11/24 - My SECRET Weapon for Social Media: AI Model Body Swap (ComfyUI) ➡️ https://youtu.be/eXkyn8mQ4fA
06/02/24 - AI Models From ANY Angle & SUPIR Upscaling (Low VRAM Method) ➡️ https://youtu.be/AyjdjF0YP_s
05/22/24 - The "Secret Sauce" to AI Model Consistency in 6 Easy Steps (ComfyUI) ➡️ https://youtu.be/nVaHinkGnDA
05/15/24 - IDM VTON on comfyUI for Realistic Clothing on Digital Models + Workflow ➡️ https://youtu.be/pFYqPf1Z7x8
05/09/24 - IDM VTON for Realistic Clothing on Digital Models + ComfyUI Workflow ➡️ https://youtu.be/WL59FqL0L-s
05/01/24 -Turn Boring Shoe into Killer Lifestyle Images with ComfyUI ➡️ https://youtu.be/ytmEQK2ILLU
04/13/24 - Wear Anything Anywhere using IPAdapter V2 (ComfyUI Tutorial) ➡️ https://youtu.be/X8UUFMuf_u4
01/21/24 - AI Revolution - Face ID v2 & Instant ID Review on (ComfyUI & Automatic 1111) ➡️ https://youtu.be/3Uj1PPwSm00
12/16/24 - How to make an AI Instagram Model Girl on ComfyUI (AI Consistent Character) ➡️ https://youtu.be/HPSu6BWTHlc
STEPS:
Creating stunning AI art with multiple characters can be frustrating when you're battling inconsistent features and duplicate faces. But fret no more! This comprehensive guide, inspired by a popular YouTube tutorial, walks you through leveraging the power of IPadapter in ComfyUI to achieve picture-perfect character consistency.
Step 1: Setting the Foundation
- Load Your Weapon of Choice: Begin by loading your preferred checkpoint model using the "Load Checkpoint" node.
- Embrace the "Anything Everywhere 3" Node: Connect the output to an "Anything Everywhere 3" node for a streamlined, organized workflow.
- Summon the IPadapter: Load an "IPadapter Advanced" node and the "IPadapter Unified Loader." Remember to download the necessary CLIP Vision and IPadapter models from the Cubiq GitHub page and place them in the correct folders within your "models" directory.
- Connect and Configure: Connect the "Unified Loader" and IPadapter outputs to the "IPadapter Advanced" inputs. Select the potent "Plus Face Portrait" model.
Step 2: Unleash the Power of Reference Images
- Load Image Batch Node: Add a "Load Image Batch from Dir" node (part of the Inspire Pack Custom Node) to link your reference images.
- Organize Your Character Folders: Create separate folders for each character, placing several clear, front-facing reference images within each.
- Connect and Point to Your Images: Connect this node to the IPadapter's image input, providing the path to your first character's folder.
Step 3: Duplicate for Multi-Character Magic
- Copy and Paste Your IPadapter Powerhouse: Duplicate the entire IPadapter node setup (hold Control, select nodes, copy with Control+C, and paste with Shift+V).
- Chain and Adjust: Connect the output of the first IPadapter to the input of the second. Replace the second IPadapter's image path with your second character's folder. Consider adjusting the weight (e.g., to 0.6) to fine-tune the model's strength.
Step 4: Wield the Power of Attention Masks
- Load Your Masks: Load attention masks to designate the areas where each character will appear in your final image. You can find pre-made masks online or create your own.
- Apply with Precision: Use a "Left" mask for the first IPadapter and a "Right" mask for the second, ensuring each IPadapter focuses on its designated character zone.
Step 5: Prompt Your Way to Perfection
- Craft Your Vision: Write a detailed positive prompt describing your desired scene, including the number of characters, their clothing, and the environment. For example, "A portrait photo of two women, one with flowing red hair and a emerald dress, the other with short black hair and a leather jacket, standing in a bustling cyberpunk city."
- Banish the Unwanted: Use a negative prompt to exclude any undesired elements.
Step 6: Dimensions, Sampling, and Decoding
- Set the Stage: Add an empty latent node to define your image dimensions (e.g., 1216x832 for landscape).
- Sampling Time: Add a K Sampler node, connecting the positive prompt, negative prompt, empty latent node, and the output from your final IPadapter. Configure your sampling settings based on your chosen checkpoint model.
- Bring Your Vision to Life: Finally, add a VAE Decoder and a "Preview/Save Image" node.
Step 7: Generate, Refine, and Repeat
- Witness the Magic: Run your workflow and marvel at your multi-character masterpiece!
- Experiment and Iterate: Don't be afraid to adjust your prompts, masks, or reference images to achieve your desired results. Play with seed values for diverse variations.
- Scale Up Your Character Roster: To add more characters, simply repeat the IPadapter duplication process, use appropriate masks (Left, Middle, Right, etc.), and update your prompts accordingly.
Congratulations! You're now equipped to conquer multi-character AI art in ComfyUI with stunning consistency.
Workflow