Speculative Sampling is proposed, a strategy that combines the modeling power of large bidirectional windows with the low-latency interaction of small ones, opening a path to serving large bidirectional models in interactive streaming systems without complex streaming post-training.
Yu-Bo Huang, Yuxuan Zhang, Jia-Ming Liu et al.· 0 citations
Audio-driven avatar generation requires realistic lip-sync, expressive motion, and real-time streaming. Recent work achieves the latter via self-forcing with Distribution Matching Distillation (DMD), but this paradigm suffers from a critical failure that has not been systematically characterized: dynamic collapse, wher...
Yubo Huang, Si-Rui Zhao, Xinchen Yao et al.· 1 citation· ⚡1
LSTMem is proposed, an LSTM-inspired online memory that equips each layer of a frozen LLM with two matrix-valued states: a cell state that accumulates history and a hidden state whose readouts correct the backbone's attention.