mengfanxu

fxmeng

·

https://fxmeng.github.io

fxmeng

AI & ML interests

None yet

Recent Activity

upvoted a paper about 8 hours ago

On the Design of Hyper-Connections: Simplification, Stability, and Scalability

updated a collection 28 days ago

updated a collection 28 days ago

View all activity

Organizations

None yet

commented a paper about 1 year ago

TPLA: Tensor Parallel Latent Attention for Efficient Disaggregated Prefill \& Decode Inference

Paper • 2508.15881 • Published Aug 21, 2025 • 10 •

commented 2 papers over 1 year ago

TransMLA: Multi-head Latent Attention Is All You Need

Paper • 2502.07864 • Published Feb 11, 2025 • 70 •

TransMLA: Multi-head Latent Attention Is All You Need

Paper • 2502.07864 • Published Feb 11, 2025 • 70 •

commented a paper almost 2 years ago

TransMLA: Multi-head Latent Attention Is All You Need

Paper • 2502.07864 • Published Feb 11, 2025 • 70 •

New activity in MMMU/MMMU almost 3 years ago

Question about "Text as Input"

#4 opened almost 3 years ago by