https://xwlu.github.io/tags/inferenceacceleration/ Recent content in InferenceAcceleration on 朝花夕拾 Hugo en-us Fri, 15 May 2026 14:16:37 +0800 https://xwlu.github.io/post/robotics/e2e/x_cache/ Thu, 30 Apr 2026 00:00:00 +0800 https://xwlu.github.io/post/robotics/e2e/x_cache/ X-Cache 发现跨生成片段的 block 残差复用，以零训练代价实现 2.6-2.7 倍世界模型推理加速