Preview for HiDream-I1 - Full

Run this workflow on InstaSD

Get started in minutes! Run this ComfyUI workflow online - no setup required.

Description

Description

Unlock the full potential of image generation with HiDream Full on ComfyUI, powered by InstaSD. This high-fidelity model leverages the advanced Diffusion Transformer (DiT) architecture to deliver stunning, photorealistic results at 50 inference steps. Ideal for creators who prioritize detail, composition, and precision, HiDream Full runs in FP16 for maximum quality and is optimized for high-VRAM GPUs.

Whether you're a digital artist, concept designer, or AI researcher, this setup offers uncompromising image generation right from your browser—no installation required.


🎯 Features

  • Photorealistic Outputs – Generate images with rich textures, clean geometry, and vibrant lighting.
  • FP16 Precision – Maintains high dynamic range and visual clarity for detailed scenes.
  • Powerful Architecture – Built on HiDream’s DiT-based diffusion for premium output quality.
  • Runs on High-End GPUs – Delivered through InstaSD’s fast, cloud-powered infrastructure.

💡 Use Cases

  • Concept Art & Visual Development – Explore character designs, environments, and visual moods.
  • Product Mockups – Create high-quality visual assets for branding or marketing.
  • AI Research & Experimentation – Benchmark and explore diffusion-based generation in FP16.

⚙️ How It Works

  1. Launch the Workflow – Load the preconfigured ComfyUI instance with HiDream Full.
  2. Customize Your Prompt – Input your creative prompt and choose generation parameters.
  3. Run & Preview – Let InstaSD handle the heavy lifting on high-VRAM GPUs.
  4. Download Results – Export high-resolution images for your projects or portfolio.

Models

FileDestinationSource
clip_g_hidream.safetensors/ComfyUI/models/text_encodersDownload
clip_l_hidream.safetensors/ComfyUI/models/text_encodersDownload
llama_3.1_8b_instruct_fp8_scaled.safetensors/ComfyUI/models/text_encodersDownload
t5xxl_fp8_e4m3fn_scaled.safetensors/ComfyUI/models/text_encodersDownload
ae.safetensors/ComfyUI/models/vaeDownload
hidream_i1_full_fp8.safetensors/ComfyUI/models/diffusion_modelsDownload

Nodes

EmptySD3LatentImageVAEDecodeMarkdownNoteCLIPTextEncodeModelSamplingSD3VAELoaderQuadrupleCLIPLoaderKSamplerUNETLoaderSaveImage