[claudesquad] update from 'lw-offload-2' on 08 Jan 26 20:53 CST

2026-01-08 20:53:08 +08:00
parent 85bcca3d17
commit a8c9f0d837
9 changed files with 894 additions and 1704 deletions
--- a/nanovllm/kvcache/init.py
+++ b/nanovllm/kvcache/init.py
@@ -76,6 +76,8 @@ def create_kvcache_manager(config: "Config") -> KVCacheManager:
        block_size=config.kvcache_block_size,
        policy=eviction_policy,
        sparse_policy=sparse_policy,
+        num_kv_buffers=getattr(config, 'num_kv_buffers', 4),
+        max_seq_len=config.max_model_len,
    )