Hello, I'm wytsai7660 (蔡維元)
As a Computer Science undergraduate, I view this field as my lifelong passion. My interests span a wide range of technologies, with a particular focus on Computer Vision. I am driven by a mission to leverage my knowledge and skills to address tangible, real-world challenges.
Research Interests
- Modern CNNs (e.g., ResNeXt, EfficientNet, ConvNeXt)
- Vision Transformers
- Vision Language Models
- Multimodal LLMs
Skills
Programming Languages
- Python
- C++
- TypeScript
- Shell Script
- LaTeX
Machine Learning & CV
- PyTorch Ecosystem
- Hugging Face Ecosystem
- OpenCV
- Scikit-learn
- Pandas & NumPy
Development & Tools
- Linux
- Git
- GitHub
- Docker
- GitHub Actions
Portfolio
In this study, a modified multimodal large language model-based approach for image quality assessment is
proposed
Repo: wytsai7660/iqa-project
Repo: wytsai7660/iqa-project
Our team earned an Honorable Mention for a multi-task classification problem on time-series table tennis
racket sensor data.
Repo: wytsai7660/AI-CUP-2025-Table-Tennis
Repo: wytsai7660/AI-CUP-2025-Table-Tennis
A website that allows citizens to propose ideas to the government, discuss each other's issues, and use AI
to simplify proposal searches
Repo: citizen-proposal-APP/CitizenProposalApp
Repo: citizen-proposal-APP/CitizenProposalApp