🙇🏻‍♂️ BRAINLESS FLUX FLOW 👨🏻‍💻

Use 🙇🏻‍♂️ BRAINLESS FLUX FLOW in native ComfyUI to create Flux.1 D text-to-image images from a Main Prompt and chosen resolution, with webp or jpeg output.

The workflow includes resolution presets, a negative prompt for SFW, native upscaling integration, FaceDetailer, LoRA loading, and Ultralytics for BBOX and SEGM models.

Enter a prompt, choose a resolution, set the scheduler and sampler, select an image extension, and press Queue.

Enter what you want to see in Main Prompt, then choose a preferred size with Resolution Size and Preset User.

Choose webp or jpeg in Image Saver Simple; the workflow also lists native upscaling integration.

Install native ComfyUI, download and extract the workflow ZIP, open it with Workflow → Open, and install custom nodes through Manager → Install Custom Nodes.

Install a Flux.1 D model, then refresh node definitions with Edit → Refresh Node Definitions or restart ComfyUI.

The listed model files are clip_g.safetensors, clip_l.safetensors, flux_vae.safetensors, t5xxl_fp8_e4m3fn.safetensors, and sam_vit_b_01ec64.pth.

The listed node packs are comfyui-image-saver, comfyui-impact-pack, comfyui-impact-subpack, comfyui-mnemic-nodes, and rgthree-comfy.

The listed folders include */models/checkpoints/ for a Flux.1 D model and */models/ultralytics/bbox/ for face_yolov8m.pt.

Set BasicScheduler steps from 12 to 32 and choose sampler_name in KSamplerSelect.

Sugestão · não verificado

For performance adjustments, bypass Ultralytics and SAMLoader, change FaceDetailer settings, or set up SageAttention yourself.

Sugestão · não verificado

Before queuing, verify that the listed model files are in their expected folders, including sam_vit_b_01ec64.pth in */models/sams/.

Sugestão · não verificado

No license is specified for the workflow.