WorkFlow for Flux.dev by Danrisi (txt2img, img2img, upscale and some QoL stuff)

This Flux.dev workflow generates images from text prompts or input images, upscales them, and saves the image outputs.

Its 34-node graph includes text encoding, image loading, sampling, VAE encoding and decoding, resizing, iterative latent upscaling, and image saving.

For text-to-image, it lists CLIPTextEncode; for image-to-image, it lists LoadImage.

SaveImage is the listed image output, with IterativeLatentUpscale and PixelKSampleUpscalerProvider listed for upscaling.

Danrisi is the listed publisher, Civitai is the source page, and Flux.1 D is the listed base model.

The downloadable archive is named workflowForFluxDevByDanrisi_v10.zip.

The listed model files are 4xNMKDSuperscale_4xNMKDSuperscale.pt, ae.safetensors, clip_l.safetensors, t5xxl_fp16.safetensors, and UltraRealistic_FineTune_Project_v4_Q8.gguf.

The listed node packs are cg-use-everywhere, comfyui_controlaltai_nodes, comfyui_essentials, comfyui-gguf, comfyui-impact-pack, controlaltai-nodes, and rgthree-comfy.

The node list includes FluxResolutionNode, FluxGuidance, ModelSamplingFlux, RescaleCFG, and KSamplerSelect.

Power Lora Loader (rgthree) is also listed among the node types.

No individual source URLs are listed for the model files, and no license is listed for the workflow; verify download sources and usage terms before use.

Sugerencia · sin verificar

Before use, verify that the package providing comfyui-gguf, the model file UltraRealistic_FineTune_Project_v4_Q8.gguf, and all listed node packs are available in your setup.

Sugerencia · sin verificar