Yuezhouhu Yuezhou Hu Github

Yuezhouhu Yuezhou Hu Github Beginner to machine learning | uc berkeley eecs. yuezhouhu has 7 repositories available. follow their code on github. I am a phd student at berkeley artificial intelligence research lab (bair), advised by prof. kurt keutzer. previously, i completed my undergrad in tsinghua university. during this time, i was fortunate to be advised by prof. jianfei chen and prof. jun zhu in tsail, and prof. tuo zhao in georgia tech.

Publications Yuezhou Hu 胡越舟 My research interests include efficient machine learning, particularly efficient training and inference. A selective knowledge distillation algorithm for efficient speculative decoders yuezhouhu adaspec. Take a look at residual context diffusion (rcd): a simple idea to boost diffusion llms—stop wasting “remasked” tokens!!! (example on aime24. rcd increases parallelism by 4x while reaching the baseline's peak accuracy.). Yuezhou hu *, harman singh*, monishwaran maheswaran*, haocheng xi, coleman hooper, jintao zhang, aditya tomar, michael w. mahoney, sewon min, mehrdad farajtabar, kurt keutzer, amir gholami, chenfeng xu.

Yue S Homepage Yue Zhou Master At Tum Take a look at residual context diffusion (rcd): a simple idea to boost diffusion llms—stop wasting “remasked” tokens!!! (example on aime24. rcd increases parallelism by 4x while reaching the baseline's peak accuracy.). Yuezhou hu *, harman singh*, monishwaran maheswaran*, haocheng xi, coleman hooper, jintao zhang, aditya tomar, michael w. mahoney, sewon min, mehrdad farajtabar, kurt keutzer, amir gholami, chenfeng xu. My research interests primarily focus on efficient machine learning, particularly efficient training and inference. recently, i am interested in:. Beginner to machine learning | uc berkeley eecs. yuezhouhu has 5 repositories available. follow their code on github. Efficient 2:4 sparse training algorithms and implementations yuezhouhu 2by4 pretrain. We propose residual context diffusion (rcd), a novel paradigm that transforms this wasted computation into a guiding signal. by treating latent representations of low confidence tokens as residual updates, rcd allows models to progressively refine their knowledge.