Ju He's picture

Ju He

turkeyju

·

https://tacju.github.io/

TACJu

AI & ML interests

None yet

Recent Activity

upvoted a paper about 19 hours ago

Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision Encoders

upvoted a paper 6 days ago

Beyond Language Modeling: An Exploration of Multimodal Pretraining

upvoted a paper 8 days ago

Compositional Generalization Requires Linear, Orthogonal Representations in Vision Embedding Models

View all activity

Organizations

commented a paper 12 months ago

FlowTok: Flowing Seamlessly Across Text and Image Tokens

Paper • 2503.10772 • Published Mar 13, 2025 • 19 •

commented a paper about 1 year ago

Democratizing Text-to-Image Masked Generative Models with Compact Text-Aware One-Dimensional Tokens

Paper • 2501.07730 • Published Jan 13, 2025 • 18 •