Hunyuan Image 2.1 Native 2K Text-to-Image + Refiner | Full ComfyUI Workflow #AIArt #ComfyUI #2K #ai

This ComfyUI workflow is for Hunyuan Image 2.1 native 2K text-to-image generation with a refiner. It takes text conditioning and can take an image, then produces decoded images for saving.

The 69-node workflow includes text encoding, latent initialization, sampling, VAE encoding and decoding, image comparison, and image saving.

The workflow includes CLIPTextEncode nodes for text conditioning and LoadImage for image input.

LoadImage and VAEEncode support image handling, while VAEDecode and SaveImage support image output.

Raretutor lists the workflow on Civitai at https://civitai.com/models/1974024; the downloadable archive is named hunyuanImage21Native2KTextToImage_v10.zip.

Listed model files include byt5_small_glyphxl_fp16.safetensors, hunyuan_image_2.1_vae_fp16.safetensors, hunyuan_image_refiner_vae_fp16.safetensors, hunyuanimage-refiner_fp8.safetensors, and hunyuanimage-refiner-v2-q8_0.gguf.

They also include hunyuanimage2.1_fp8_e4m3fn.safetensors, hunyuanimage2.1-distilled-v2-q8_0.gguf, pig_hunyuan_image_refiner_vae_fp32-f16.gguf, pig_hunyuan_image_vae_fp32-f16.gguf, qwen_2.5_vl_7b_fp8_scaled.safetensors, and Qwen2.5-VL-7B-Instruct-Q4_K_M.gguf.

The workflow lists these node packs: comfyui-kjnodes, comfyui-multigpu, gguf, rgthree-comfy, and wavespeed.

The listed node-pack pins are comfyui-kjnodes 1.1.4, comfyui-multigpu 803cf542d9d517aae769d07f8bd4e1f5a9dda80a, gguf 2.4.7, rgthree-comfy 1.0.2509062212, and wavespeed 16ec6f344f8cecbbf006d374043f85af22b7a51d.

Before use, verify that the listed model files are available and that the node packs match the listed pins.

Vorschlag · nicht geprüft

No model-use license is specified for this workflow; verify the applicable terms before use.

Vorschlag · nicht geprüft