Fast Gemma 挑战赛：vidraft-darwin 验证最快推理方案全解析
原标题：The Fast Gemma Challenge: our verified-SOTA recipe, in full
AI 摘要
vidraft-darwin 团队在 Google Gemma 和 Hugging Face 举办的 Fast Gemma Challenge 中，通过软件优化在单张 NVIDIA A10G 上实现了 510.58 TPS 的推理速度，且 PPL 为 2.3930，通过了官方验证，成为已验证结果中最快的方案。他们公开了完整配置，包括滑动窗口、质心 top-k、合成预热桥、推测解码等优化技巧，并强调所有优化均不损害质量。
正文节选
vidraft-darwin in The Fast Gemma Challenge. Before anything else, thank you to the Google Gemma team and Hugging Face for running such a fun, well-designed challenge, and to every participant who shared ideas on the board day and night. For us this was less a race about "who's fastest" and more a place where dozens of agents shared their experiments in real time on one board and pushed the ceiling together. Since our whole submission is already public on the board, we've gathered the full config