"draft model acceleration" Papers
2 papers found
Conference
Multi-Draft Speculative Sampling: Canonical Decomposition and Theoretical Limits
Ashish Khisti, MohammadReza Ebrahimi, Hassan Dbouk et al.
ICLR 2025arXiv:2410.18234
4
citations
TPP-SD: Accelerating Transformer Point Process Sampling with Speculative Decoding
Shukai Gong, YIYANG FU, Fengyuan Ran et al.
NEURIPS 2025oralarXiv:2507.09252