Pulid 2 is here to revolutionize AI image generation. With dual image support, Tea Cache integration, and the speed-boosting Wave Speed feature, this update makes creating stunning visuals faster and easier. Let’s dive in and explore how to set up Pulid 2, optimize workflows, and make the most of its cutting-edge tools.
Models Require to Run Workflow
Follow these steps to get Pulid 2 up and running:
Flux Model Download Link
- 32bit/16bit (~22GB VRAM): model, encoder
- 8bit gguf (~12GB VRAM): model, encoder
- 8 bit FP8 e5m2 (~12GB VRAM): model, encoder
- 8 bit FP8 e4m3fn (~12GB VRAM): model, encoder
- Clip and VAE (for all models): clip, vae
1. Remove Conflicting Nodes
Uninstall older versions, like ComfyUI-PuLID-Flux, to avoid conflicts. These older nodes share the same name (ApplyPulidFlux).
2. Install Pulid 2
- Open the ComfyUI Manager section.
- Search for “Pulid 2” and install it.
3. Fix Missing Nodes
- After uploading the workflow, check for missing nodes in the “Manager Missing Nodes” section.
- Install required nodes directly from there.
4. Install Dependencies
- Navigate to the ComfyUI_PuLID_Flux_ll folder using CMD.
- Run this command:
pip install -r requirements.txt
5. Download Required Models
Save these files in the correct directories:
- Download pulid_flux_v0.9.1.safetensors:
/models/pulid/ - EVA02-CLIP-L-14-336:
/ComfyUI/models/clip/(Auto-download supported) - AntelopeV2 Models:
/ComfyUI/models/insightface/models/antelopev2/(Auto-download supported) - Parsing Models:
Save parsing models in /ComfyUI/models/facexlib/.
7. Add Tea Cache and Wave Speed Nodes
Once installed, your setup is complete and ready to go!
WAVE SPEED
git clone https://github.com/chengzeyi/Comfy-WaveSpeed.git