fix: Minimize gpu memory fragmentation

qdrant · Feb 28, 2025 · ed5bc00 · ed5bc00
1 parent 2082108
commit ed5bc00
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/fastembed/common/onnx_model.py b/fastembed/common/onnx_model.py
@@ -85,7 +85,7 @@ def _load_onnx_model(
 
         so = ort.SessionOptions()
         so.graph_optimization_level = ort.GraphOptimizationLevel.ORT_ENABLE_ALL
-
+        so.add_session_config_entry("memory.enable_memory_arena_shrinkage", "1")
         if threads is not None:
             so.intra_op_num_threads = threads
             so.inter_op_num_threads = threads