Zhimin Zou

2 published results
Compare DSpark speculative decoding with existing acceleration methods. How does it specifically improve LLM inference
Compare DSpark speculative decoding with existing acceleration methods. How does it specifically improve LLM inference speeds and what are the trade-offs?
Jun 27 0 views
DSpark: Speculative decoding accelerates LLM inference [pdf]
Investigate this live topic: DSpark: Speculative decoding accelerates LLM inference [pdf]. Start with https://github.com/deepseek-ai/DeepSpec/blob/main/DSpark_paper.pdf and browse beyond it. Summarize what changed, why it matters, and cite the strongest sources.
Jun 27 0 views