【Anima】Simple T2I Workflow with Upscale, Detailers and ControlNet — v3.0

Use this Anima v3.0 text-to-image workflow to turn a text prompt into images, with optional Anima-LLLite ControlNet, face detailing, and Ultimate SD Upscale; it includes image preview, comparison, and saving.

Its nodes cover prompt encoding and image sampling, Anima-LLLite application, face detailing, upscaling, image comparison, preview, and saving.

Enter a text prompt; the workflow includes image preview and save steps, plus image comparison.

The description points to the `ComfyUI-Anima-LLLite` custom node for the ControlNet.

Listed ComfyUI node packs and versions: `comfyui_ultimatesdupscale` 1.6.2, `comfyui-impact-pack` 8.28.2, `comfyui-impact-subpack` 1.3.5, and `rgthree-comfy` 1.0.2512112053.

Model files named for the workflow include `4x_foolhardy_Remacri.pth`, `anima-highres-aesthetic-boost.safetensors`, `anima-lllite-any-test-like-1-step2000.safetensors`, `anima-preview3-base.safetensors`, and `bbox/face_yolov8n.pt`.

Also named: `bbox/hand_yolov8s.pt`, `qwen_3_06b_base.safetensors`, `qwen_image_vae.safetensors`, `sam_vit_b_01ec64.pth`, and `segm/PitHandDetailer-v2-Test-v9c.pt`.

You can bypass the Anima-LLLite ControlNet.

Before enabling ControlNet, check which file your setup should use: the v3.0 workflow names `anima-lllite-any-test-like-1-step2000.safetensors`, while the description recommends `anima-lllite-any-test-like-v2.safetensors`.

Vorschlag · nicht geprüft