Rongfu Bai

I’m Rongfu Bai 白镕辅, a senior student at the School of Software Engineering, Xi’an Jiaotong University. I joined the ENCODE lab at Westlake University as a visiting student advised by Prof. Huan Wang in the summer of 2025.

I mainly work on Computer Vision, Model Compression, and Unsupervised Learning. I am also deeply interested in LLMs/VLMs.

Experience

I worked on instance segmentation of PolSAR (Polarimetric Synthetic Aperture Radar) images via contrastive learning under the guidance of Professor Haixia Bi in my sophomore year, which sparked my interest in Computer Vision and Unsupervised Learning.

At ENCODE Lab, I focused on accelerating inference for quantized models, gaining hands-on experience with advanced quantization techniques and CUDA programming. I contributed to a project that ternarized the linear layers of the VMamba model and replaced conventional convolution operations with bitwise computations, drastically reducing the number of multiply–accumulate operations. This innovation significantly lowered energy consumption for inference on edge devices and offered a theoretically substantial speedup, demonstrating a practical path toward efficient deployment of deep models in resource-constrained environments.