How to Run HiDream on ComfyUI | EASY Setup Guide
Added 2025-04-22 18:25:05 +0000 UTCIn this tutorial, I walk you through how to run the open-source HiDream image generation model on ComfyUI, including support for both full and quantized versions. Whether you're using a powerful GPU or working with limited VRAM, this guide is designed to help you get the most out of HiDream's advanced capabilities.
HiDream Gradio & model comparison: https://www.patreon.com/posts/126749471
What You’ll Learn in This Video:
✅The differences between full and quantized versions of HiDream (Q8, Q6, Q2, etc.)
✅How to set up HiDream with ComfyUI
✅Folder structure and placement of model files
✅Explanation of quadruple CLIP loaders and HiDream-specific encoders
✅How to enhance image quality using the Detail Daemon node
✅Tips for working within limited VRAM (8GB+ GPUs)
✅Workflow optimization tricks and model sampling observations
✅Troubleshooting and VRAM-saving techniques for RTX 3090 & similar GPUs
📂Download Links & Resources:
▸HiDream GGUF - https://huggingface.co/city96/HiDream-I1-Full-gguf/tree/main
▸HiDream bf16 & fp8 safetensors - https://huggingface.co/Comfy-Org/HiDream-I1_ComfyUI/tree/main/split_files/diffusion_models
▸HiDreamTextEncoders -https://huggingface.co/Comfy-Org/HiDream-I1_ComfyUI/tree/main/split_files/text_encoders
▸Hidream Vae - https://huggingface.co/Comfy-Org/HiDream-I1_ComfyUI/tree/main/split_files/vae
▸Image Generator: https://kaijugen.com/
▸Prompt Database: https://promptcrafters.co/
▸ Sampler comparison: [coming soon]
▸ DetailDaemon Video: [coming soon]
