🔥 HiDream: Advanced AI Image Generation in Multiple Formats

📌 Overview

HiDream is now provided in several AI model versions and formats (like .safetensors and .gguf), each suited for different needs regarding speed, quality, VRAM usage, and software compatibility. Understanding these helps you select the best option for your hardware and find the right download links.

🔑 This Guide Covers

1. Core Models: The main variants (Full, Dev, Fast) and where to find them

2. Required Components: The Text Encoder and VAE models needed for complete functionality

3. Quantized Versions: The GGUF model versions and their sources

4. Hardware Requirements: General GPU VRAM requirements for these models

⚡ Core HiDream Image Models

• 1. HiDream Full

  • Focus: Highest Quality & Detail

  • Settings: Steps: 50, Sampler: uni_pc, Scheduler: simple, CFG: 5.0 (Needs Negative Prompt)

• 2. HiDream Dev

  • Focus: Speed & Experimentation

  • Settings: Steps: 28, Sampler: lcm, Scheduler: normal, CFG: 1.0 (No Negative Prompt)

• 3. HiDream Fast

  • Focus: Maximum Speed

  • Settings: Steps: 16, Sampler: lcm, Scheduler: normal, CFG: 1.0 (No Negative Prompt)

Download Location (Split FP16/FP8 Diffusion Models): Hugging Face Repository

Note: Download all parts for the desired model (e.g., hidream_i1_full_fp16.safetensors) and place them in the ComfyUI/models/diffusion_models/ folder.

🧠 Text Encoder & VAE Models

These components are essential for interpreting your prompts and finalizing the image. They need to be loaded alongside the main image models, consuming additional VRAM.

Required HiDream Text Encoders

1. clip_g_hidream.safetensors (1.39 GB)

2. clip_l_hidream.safetensors (248 MB)

3. llama_3.1_8b_instruct_fp8_scaled.safetensors (9.08 GB)

4. t5xxl_fp8_e4m3fn_scaled.safetensors (5.16 GB)

Download Location (Split Text Encoders): Hugging Face Repository

Note: Download all parts for the required encoders (clip_g, clip_l, t5xxl, llama3.1) and place them in the ComfyUI/models/text_encoders/ folder.

VAE

• ae.safetensors

📁 Recommended Folder Structure

For optimal organization of HiDream models and components, follow this recommended structure:

📂 ComfyUI/
├── 📂 models/
│   ├── 📂 diffusion_models/
│   │   └── 📄 hidream_i1_[variant].safetensors (or .gguf)
│   ├── 📂 text_encoders/
│   │   ├── 📄 clip_g_hidream.safetensors
│   │   ├── 📄 clip_l_hidream.safetensors
│   │   ├── 📄 llama_3.1_8b_instruct_fp8_scaled.safetensors
│   │   └── 📄 t5xxl_fp8_e4m3fn_scaled.safetensors
│   ├── 📂 vae/
│   │   └── 📄 ae.safetensors

📊 GGUF Model Versions (Quantized HiDream)

Available HiDream-I1 GGUF Quantizations & Download Links

1. HiDream Full GGUF

   • https://huggingface.co/city96/HiDream-I1-Full-gguf

2. HiDream Fast GGUF

   • https://huggingface.co/city96/HiDream-I1-Fast-gguf

3. HiDream Dev GGUF

   • https://huggingface.co/city96/HiDream-I1-Dev-gguf

Note: K-quants (_K_M, KS) often offer better quality than older standard quants (_0, _1) at similar bit levels.

💻 GPU VRAM Requirements & Model Suitability

VRAM is crucial. You need enough space for the model, necessary auxiliaries (encoders, VAE), software overhead, and the generation process.

General GGUF Suitability by GPU VRAM (Focusing on HiDream Full GGUF sizes)

1. 8GB VRAM

• Best suited for Q2_K (6.56 GB) or potentially Q3_K_S (8.21 GB) / Q3_K_M (8.77 GB) with some offloading.

2. 12GB VRAM (e.g., RTX 3060 12GB)

• Can comfortably run Q3 levels fully in VRAM.

• Can run Q4 levels (up to Q4_K_M @ 11.5 GB), but the larger Q4 models will likely require some layers offloaded to CPU/RAM.

• Q5 levels (up to 13.5 GB) are possible but require more significant offloading.

• Recommendation: Start with Q4_K_M or Q4_K_S; try Q3_K_M if speed is an issue due to offloading.

3. 16GB VRAM

• Can likely run Q5 levels (up to 13.5 GB) fully in VRAM.

• Can run Q6_K (14.7 GB), possibly needing minimal offload.

4. 24GB VRAM (or more)

• Can comfortably run Q6_K (14.7 GB) and likely higher quantizations like Q8_0.

• The full FP16 (34.2 GB) requires substantial offloading even on 24GB, better suited for 48GB+ cards or multi-GPU setups if aiming for full VRAM utilization.

🙏 Credits

• Special thanks to Black Forest Labs for developing the original FLUX.1-Fill-dev model.

• Big thanks to city96 for pioneering the GGUF journey! 🙌

👨‍💻 Developer Information

This guide was created by Abdallah Al-Swaiti:

1. Hugging Face

2. GitHub

3. LinkedIn

For additional tools and updates, check out my ComfyUI-OllamaGemini

✨ Elevate Your Creative Vision with HiDream ✨