Black Mixture
patreon
🔥 (New Release) Black Mixture Professional Watermark Removal ComfyUI Workflow
🕑 Added 2024-12-05 06:45:51 +0000 UTCHey Black Mixture Fam! We're excited to drop our latest workflow that harnesses the power of Black Forest Labs' Flux Fill model for professional-grade watermark removal. We've optimized all settings for the best possible results, making it easy to achieve clean, high-quality outputs. Stay tuned for our upcoming tutorial video!
📦 Required Models:
Main Checkpoint:
Flux Fill Model: https://huggingface.co/black-forest-labs/FLUX.1-Fill-dev/tree/main
Place in: 📂ComfyUI\models\diffusion_models
CLIP Model:
Flux Text Encoders: huggingface.co/comfyanonymous/flux_text_encoders
Place in: 📂ComfyUI\models\clip
VAE Model:
ae.safetensors: https://huggingface.co/black-forest-labs/FLUX.1-Fill-dev/blob/main/ae.safetensors
Place in: 📂ComfyUI\models\vae
✨ Features:
Optimized settings for professional results
Intuitive color-coded sections
Simple workflow layout
Comprehensive watermark removal
Compatible with latest ComfyUI
No visible node wires
🎨 Workflow Sections:
🟥 Load Models: Set up your checkpoints and components
🟨 Image Input: Load and mask your watermarked image
🟦 Prompting: Fine-tune your generation parameters
🟩 Generation: Control settings and compare results
🟪 Info: Version and support details
⚙️ Quick Start:
1. Download required models
2. Place in correct folders as specified above
3. Load workflow into ComfyUI
4. Hit "Install Missing Custom Nodes" if prompted
5. Load your image and right-click → open in mask editor
6. Draw rough masks over the watermarks → save to node
7. Hit Queue to start removing watermarks!
💡 Need help? Check the comments or join our community Discord.
📹 Tutorial video is here! In this video we show you how to get the most out of this workflow. https://youtu.be/58ybYm4jzvg
Comments
Black Mixture
This error is related to having an outdated version of ComfyUI. The older versions don't support the latest Flux Fill model. Updating to the latest stable build of ComfyUI will fix this specific error.
Alexander Pereira
This is dope.. potentially, cause I got an error, any idea what it might be? I downloaded the Flux Fill model and VAE and loaded them up, painted in the mask, didn't enter any prompt, everything else is as-is. I even used the very same shutterstock image you used, resolution 1409x1600 px. The error: Error(s) in loading state_dict for Flux: size mismatch for img_in.weight: copying a param with shape torch.Size([3072, 384]) from checkpoint, the shape in current model is torch.Size([3072, 64]). Full error message: https://pastebin.com/143kBj2x
