Skip to main content
Run any Skill in Manus
with one click

model-infer-kvcache

Stars12
Forks3
UpdatedJune 18, 2026 at 02:56

基于 PyTorch 框架的昇腾 NPU 模型推理 KVCache 优化技能。分析并改造 LLM 推理模型的 KVCache 实现,覆盖 Legacy 连续缓存与分页注意力(Paged Attention)配 FA 融合算子、MLA 压缩缓存、SlidingWindow / 多 attn_type 混合。触发场景:KVCache 管理实现、分页注意力接入、KV 压缩、FA 融合算子、OOM / 性能问题、block_table / slot_mapping 构造。支持框架部署与独立部署两种模式。

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

File Explorer
4 files
SKILL.md
readonly