AI Model Training & Workflow R&D (Independent Projects) | Original Creator on Liblib Platform
Managed end-to-end LoRA fine-tuning for image-based AI models, including dataset collection and high-precision manual tagging. Developed and standardized automated workflows using ComfyUI to increase efficiency in data labeling and model training. Applied controlled-variable experiments to optimize labeling consistency and output stability. • Curated and tagged images for LoRA dataset creation • Built automated visual labeling pipelines in ComfyUI • Conducted iterative reviews to refine model accuracy • Ensured labels met professional visual and commercial standards