Merge pull request #27 from codefuse-ai/modelcache_dev_mm

Modelcache for mm

Author: peng3307165

.gitignore

@@ -135,5 +135,11 @@ dmypy.json
 /embedding_npy
 /flask_server
 *.bin
+**/maya_embedding_service
+
+*.ini
+
+**/multicache_serving.py
 **/modelcache_serving.py
-*.ini
+
+**/model/

flask4modelcache.py

@@ -177,7 +177,6 @@ def user_backend():
         return json.dumps(result)
 
     if request_type == 'register':
-        # iat_type = param_dict.get("iat_type")
         response = adapter.ChatCompletion.create_register(
             model=model
         )

model/clip_zh/__init__.py

@@ -0,0 +1,12 @@
+# -*- coding: utf-8 -*-
+"""
+   Alipay.com Inc.
+   Copyright (c) 2004-2023 All Rights Reserved.
+   ------------------------------------------------------
+   File Name : __init__.py.py
+   Author : fuhui.phe
+   Create Time : 2024/5/7 14:05
+   Description : description what the main function of this file
+   Change Activity: 
+        version0 : 2024/5/7 14:05 by fuhui.phe  init
+"""

modelcache/adapter/adapter_query.py

@@ -30,17 +30,20 @@ def adapt_query(cache_data_convert, *args, **kwargs):
             report_func=chat_cache.report.embedding,
         )(pre_embedding_data)
 
+        # print('embedding_data: {}'.format(embedding_data))
+
     if cache_enable:
         cache_data_list = time_cal(
             chat_cache.data_manager.search,
-            func_name="milvus_search",
+            func_name="vector_search",
             report_func=chat_cache.report.search,
         )(
             embedding_data,
             extra_param=context.get("search_func", None),
             top_k=kwargs.pop("top_k", -1),
             model=model
         )
+        print('cache_data_list: {}'.format(cache_data_list))
         cache_answers = []
         cache_questions = []
         cache_ids = []
@@ -78,8 +81,8 @@ def adapt_query(cache_data_convert, *args, **kwargs):
             return
 
         for cache_data in cache_data_list:
+            print('cache_data: {}'.format(cache_data))
             primary_id = cache_data[1]
-            start_time = time.time()
             ret = chat_cache.data_manager.get_scalar_data(
                 cache_data, extra_param=context.get("get_scalar_data", None)
             )

modelcache/core.py

@@ -4,7 +4,7 @@
 from modelcache.processor.post import first
 from modelcache.similarity_evaluation import ExactMatchEvaluation
 from modelcache.similarity_evaluation import SimilarityEvaluation
-from modelcache.embedding.string import to_embeddings as string_embedding
+from modelcache.embedding.string_text import to_embeddings as string_embedding
 from modelcache.report import Report
 from modelcache.config import Config
 from modelcache.utils.cache_func import cache_all
@@ -19,7 +19,10 @@ def __init__(self):
         self.cache_enable_func = None
         self.query_pre_embedding_func = None
         self.insert_pre_embedding_func = None
+        self.mm_query_pre_embedding_func = None
+        self.mm_insert_pre_embedding_func = None
         self.embedding_func = None
+        self.embedding_concurrent_func = None
         self.data_manager: Optional[DataManager] = None
         self.similarity_evaluation: Optional[SimilarityEvaluation] = None
         self.post_process_messages_func = None

modelcache/embedding/string.py modelcache/embedding/string_text.pymodelcache/embedding/string.py renamed to modelcache/embedding/string_text.py

@@ -212,6 +212,7 @@ def import_data(
             cache_datas.append([ans, question, embedding_data, model])
 
         ids = self.s.batch_insert(cache_datas)
+        print('ids: {}'.format(ids))
         logging.info('ids: {}'.format(ids))
         self.v.mul_add(
             [

modelcache/embedding/timm.py modelcache/embedding/timm_embedding.pymodelcache/embedding/timm.py renamed to modelcache/embedding/timm_embedding.py

@@ -94,10 +94,6 @@ class CacheStorage(metaclass=ABCMeta):
     def create(self):
         pass
 
-    @abstractmethod
-    def batch_insert(self, all_data: List[CacheData]):
-        pass
-
     @abstractmethod
     def insert_query_resp(self, query_resp, **kwargs):
         pass

modelcache/manager/data_manager.py

@@ -136,6 +136,7 @@ def _get_collection(self, collection_name):
 
     def mul_add(self, datas: List[VectorData], model=None):
         collection_name_model = self.collection_name + '_' + model
+        print('collection_name_model: {}'.format(collection_name_model))
         self._create_collection(collection_name_model)
 
         data_array, id_array = map(list, zip(*((data.data, data.id) for data in datas)))
@@ -147,6 +148,7 @@ def search(self, data: np.ndarray, top_k: int = -1, model=None):
         if top_k == -1:
             top_k = self.top_k
         collection_name_model = self.collection_name + '_' + model
+        print('collection_name_model: {}'.format(collection_name_model))
         self._create_collection(collection_name_model)
         search_result = self.col.search(
             data=data.reshape(1, -1).tolist(),