update

app.py

@@ -51,6 +51,7 @@ fuser_config.model_name = "llm-blender/gen_fuser_3b" # our pre-trained fuser
 fuser_config.max_length = 1024
 fuser_config.candidate_maxlength = DEFAULT_CANDIDATE_MAX_LENGTH
 blender_config = llm_blender.BlenderConfig()
+blender_config.load_in_8bit = True
 blender_config.device = "cuda" # blender ranker and fuser device
 blender = llm_blender.Blender(blender_config, ranker_config, fuser_config)