[LLM] First commit the llm deployment code

scripts/merge_cache_scale.py

@@ -0,0 +1,51 @@
+"""
+# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""
+
+import os
+import json
+import argparse
+import numpy as np
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--model_dir", type=str, required=True)
+args = parser.parse_args()
+ep_num = 4
+model_dir = args.model_dir
+
+scale_dicts = []
+
+for i in range(ep_num):
+    scale_dict = {}
+    path = os.path.join(model_dir, f"cachekv_scales_{i}.json")
+    with open(path, "r") as file:
+        scale_c = json.load(file)
+        for k, v in scale_c.items():
+            scale_dict[k] = v
+    scale_dicts.append(scale_dict)
+
+new_dict = {}
+for k in scale_dicts[0].keys():
+    v_list = [scale_dicts[i][k] for i in range(ep_num)]
+    v = np.concatenate(v_list, axis=1)
+    v = v.tolist()
+    new_dict[k] = v
+
+res_file = os.path.join(model_dir, "cachekv_scales.json")
+
+b = json.dumps(new_dict)
+f = open(res_file, "w")
+f.write(b)
+f.close()