Used for generating full-body and upper-body sample images with trained LoRA using various prompt styles and seed settings. Ensured high-quality anatomy representation, especially for hands and proportions.

Used to refine specific poses and hand details in generated outputs by reprocessing base images with controlled denoising strength. Helped improve consistency in tricky anatomy areas like fingers and feet.

Used for correcting missing limbs, distorted fingers, or background artifacts. Essential for cleaning up and improving selected outputs during LoRA testing.

Used to build structured generation pipelines in ComfyUI and A1111. Compared different sampling settings and LoRA epochs to optimize anatomy fidelity and style coherence.

Used to test motion consistency of LoRA in video-based outputs. Ensured that generated character hands and proportions remain stable across frames.

Used for prompt-based video generation tests with animated sequences to see if LoRA can maintain consistent hand and body anatomy in motion.

Used to convert LoRA-generated images into short animations and evaluate visual consistency. Helped identify any recurring distortions in limb movement or hand gestures.

Used to stabilize poses and enforce body structure using pose, depth, and openpose modules. Helped preserve correct limb positioning and enhance hand gestures in tricky compositions.