From 8ec0edcfe3471c6f2db1ed726ca251ca18919d6a Mon Sep 17 00:00:00 2001
From: OlivierDehaene <23298448+OlivierDehaene@users.noreply.github.com>
Date: Fri, 30 Jun 2023 16:47:00 +0200
Subject: [PATCH] fix

---
 server/text_generation_server/models/flash_causal_lm.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/server/text_generation_server/models/flash_causal_lm.py b/server/text_generation_server/models/flash_causal_lm.py
index e932b5f1..cb23bfcd 100644
--- a/server/text_generation_server/models/flash_causal_lm.py
+++ b/server/text_generation_server/models/flash_causal_lm.py
@@ -94,6 +94,7 @@ class CacheManager:
             slots.append(allocated_slots)
             block_tables.append(allocated_blocks.tolist())
             block_tables_tensor[i, :needed_blocks] = allocated_blocks
+            cumulative_blocks += needed_blocks
 
         batch.needed_blocks_slots = None
         batch.block_tables = block_tables