Kohaku BlueLeaf
Shih-Ying Yeh 葉適穎
CS master's student at National Tsing Hua University in Taiwan. Published at ICLR 2024 (LyCORIS) and ICLR 2026 (TIPO).
I research and build AI systems end-to-end — from neural network training and GPU kernels to infrastructure, databases, RAG frameworks, and agent systems.
Leading Kohaku Lab, a virtual open-source research lab. Currently exploring VAE/AE/VQ-VAE architectures for generative modeling, building agent frameworks (KohakuTerrarium), and playing with AI board game engines. Everything we build is 100% open access.
"AI art should look like AI, not humans."
Recent Projects
LyCORIS
Lora beYond Conventional methods. Comprehensive PEFT library for neural networks, published at ICLR 2024.
TIPO
Text to Image with Text Presampling for Optimal Prompting. Published at ICLR 2026.
KohakuHub
Self-hosted HuggingFace alternative, fully compatible with HuggingFace Hub, Transformers and Diffusers.
KohakuRiver
Lightweight cluster manager that turns a small fleet of nodes into one powerful computer using Docker.
KohakuBoardGame
Board game engine with NNUE neural network evaluation and PVS search. Supports MiniChess, MiniShogi, and Gomoku.
KohakuTerrarium
Async-first Python agent framework. YAML-driven config with nested sub-agents and channel-based coordination.
Recent Publications
TIPO: Text to Image with Text Presampling for Prompt Optimization
ImagenWorld: Stress-Testing Image Generation Models with Explainable Human Evaluation on Open-ended Real-World Tasks
Recent Career
Open Source Contributor
@ Kohaku LabDeveloped and maintained open-source projects related to computer vision and AI. Included personal project such as LyCORIS, TIPO, PixelOE, HakuLatent and more, and community projects such as stable-diffusion-webui, sd-scripts and more.
ML/NN Engineer (Part-time)
@ Comfy OrgCollaborated with comfy team to improve the usability and performance of Comfy UI eco system. Mainly working on lora-related system such as lora loader, weight patcher and trainer.
ML Engineer
@ AppleCollaborate with research team in Apple Inc. to explore the possibility of new form dataset. The works: "Graph-Based Captioning: Enhancing Visual Descriptions by Interconnecting Region Captions" have published on Arxiv and Apple official github.
