[claudesquad] update from 'int-minference-1' on 08 Jan 26 23:22 CST

2026-01-08 23:22:38 +08:00
parent 0bfe1984ef
commit ea4e904de0
11 changed files with 853 additions and 533 deletions
--- a/nanovllm/kvcache/sparse/full_policy.py
+++ b/nanovllm/kvcache/sparse/full_policy.py
@@ -25,6 +25,7 @@ class FullAttentionPolicy(SparsePolicy):
    # Full attention supports both prefill and decode
    supports_prefill = True
    supports_decode = True
+    requires_block_selection = False  # Load all blocks, no selective loading

    def select_blocks(
        self,