diff --git a/invokeai/backend/model_manager/load/model_cache/model_locker.py b/invokeai/backend/model_manager/load/model_cache/model_locker.py
index a275987773..269ac60479 100644
--- a/invokeai/backend/model_manager/load/model_cache/model_locker.py
+++ b/invokeai/backend/model_manager/load/model_cache/model_locker.py
@@ -60,5 +60,5 @@ class ModelLocker(ModelLockerBase):
 
         self._cache_entry.unlock()
         if not self._cache.lazy_offloading:
-            self._cache.offload_unlocked_models(self._cache_entry.size)
+            self._cache.offload_unlocked_models(0)
             self._cache.print_cuda_stats()