Developed LoRA support for WanModel within the ModelTC/LightX2V repository, focusing on parameter-efficient fine-tuning for deep learning workflows. The work involved designing a dedicated LoRA wrapper and integrating a new module to handle LoRA weight loading, enabling runtime application of LoRA weights through enhanced script argument parsing and model loading logic. By implementing these features in Python and leveraging expertise in model adaptation and machine learning, the developer established a foundation for faster experimentation and reduced computational requirements when working with WanModel variants. This contribution addressed the need for efficient model adaptation without introducing additional bug fixes during the period.