Merge pull request #679 from Ikaros-521/owner

通义千问补充大模型常用配置项和联网搜索功能
Ikaros-521 · Mar 1, 2024 · 049837f · 049837f
2 parents f5df072 + 6bedadf
commit 049837f
Show file tree

Hide file tree

Showing 5 changed files with 40 additions and 2 deletions.
diff --git a/config.json b/config.json
@@ -313,6 +313,11 @@
     "model": "qwen-max",
     "preset": "你是一个专业的虚拟主播",
     "api_key": "",
+    "temperature": 0.9,
+    "top_p": 1.0,
+    "top_k": 3,
+    "enable_search": true,
+    "max_tokens": 4096,
     "history_enable": true,
     "history_max_len": 300
   },

diff --git a/config.json.bak b/config.json.bak
@@ -313,6 +313,11 @@
     "model": "qwen-max",
     "preset": "你是一个专业的虚拟主播",
     "api_key": "",
+    "temperature": 0.9,
+    "top_p": 1.0,
+    "top_k": 3,
+    "enable_search": true,
+    "max_tokens": 4096,
     "history_enable": true,
     "history_max_len": 300
   },

diff --git a/tests/test_tongyi/tongyi.py b/tests/test_tongyi/tongyi.py
@@ -88,6 +88,11 @@ def get_resp(self, prompt):
                     self.config_data['model'],
                     messages=messages,
                     result_format='message',  # set the result to be "message" format.
+                    temperature=self.config_data['temperature'],
+                    top_p=self.config_data['top_p'],
+                    top_k=self.config_data['top_k'],
+                    enable_search=self.config_data['enable_search'],
+                    max_tokens=self.config_data['max_tokens'],
                 )
                 if response.status_code == HTTPStatus.OK:
                     logging.debug(response)
@@ -131,6 +136,11 @@ def get_resp(self, prompt):
         "model": "qwen-max",
         "preset": "你是一个专业的虚拟主播",
         "api_key": "sk-",
+        "temperature": 0.9,
+        "top_p": 0.9,
+        "top_k": 3,
+        "enable_search": True,
+        "max_tokens": 1024,
         "history_enable": True,
         "history_max_len": 20,
     }

diff --git a/utils/gpt_model/tongyi.py b/utils/gpt_model/tongyi.py
@@ -88,6 +88,11 @@ def get_resp(self, prompt):
                     self.config_data['model'],
                     messages=messages,
                     result_format='message',  # set the result to be "message" format.
+                    temperature=self.config_data['temperature'],
+                    top_p=self.config_data['top_p'],
+                    top_k=self.config_data['top_k'],
+                    enable_search=self.config_data['enable_search'],
+                    max_tokens=self.config_data['max_tokens'],
                 )
                 if response.status_code == HTTPStatus.OK:
                     logging.debug(response)
@@ -131,6 +136,11 @@ def get_resp(self, prompt):
         "model": "qwen-max",
         "preset": "你是一个专业的虚拟主播",
         "api_key": "sk-",
+        "temperature": 0.9,
+        "top_p": 0.9,
+        "top_k": 3,
+        "enable_search": True,
+        "max_tokens": 1024,
         "history_enable": True,
         "history_max_len": 20,
     }

diff --git a/webui.py b/webui.py
@@ -1168,6 +1168,10 @@ def common_textarea_handle(content):
                     config_data["tongyi"]["api_key"] = input_tongyi_api_key.value
                     config_data["tongyi"]["model"] = select_tongyi_model.value
                     config_data["tongyi"]["preset"] = input_tongyi_preset.value
+                    config_data["tongyi"]["temperature"] = round(float(input_tongyi_temperature.value), 2)
+                    config_data["tongyi"]["top_p"] = round(float(input_tongyi_top_p.value), 2)
+                    config_data["tongyi"]["top_k"] = int(input_tongyi_top_k.value)
+                    config_data["tongyi"]["enable_search"] = switch_tongyi_enable_search.value
                     config_data["tongyi"]["history_enable"] = switch_tongyi_history_enable.value
                     config_data["tongyi"]["history_max_len"] = int(input_tongyi_history_max_len.value)
 
@@ -2800,8 +2804,12 @@ def common_textarea_handle(content):
                             value=config.get("tongyi", "model")
                         ).style("width:150px")
                         input_tongyi_api_key = ui.input(label='密钥', value=config.get("tongyi", "api_key"), placeholder='API类型下，DashScope平台申请的API密钥')
-                        input_tongyi_preset = ui.input(label='预设', placeholder='API类型下，用于指定一组预定义的设置，以便模型更好地适应特定的对话场景。', value=config.get("tongyi", "preset")).style("width:600px") 
-
+                        input_tongyi_preset = ui.input(label='预设', placeholder='API类型下，用于指定一组预定义的设置，以便模型更好地适应特定的对话场景。', value=config.get("tongyi", "preset")).style("width:500px") 
+                        input_tongyi_temperature = ui.input(label='temperature', value=config.get("tongyi", "temperature"), placeholder='控制输出的随机性。').style("width:100px")
+                        input_tongyi_top_p = ui.input(label='top_p', value=config.get("tongyi", "top_p"), placeholder='在抽样时考虑的标记的最大累积概率。根据其分配的概率对标记进行排序，以仅考虑最可能的标记。Top-k采样直接限制要考虑的标记的最大数量，而Nucleus采样则基于累积概率限制标记的数量。').style("width:100px")
+                        input_tongyi_top_k = ui.input(label='top_k', value=config.get("tongyi", "top_k"), placeholder='在抽样时考虑的标记的最大数量。Top-k采样考虑一组top_k最有可能的标记。默认值为40。').style("width:100px")
+                        switch_tongyi_enable_search = ui.switch('联网搜索', value=config.get("tongyi", "enable_search")).style(switch_internal_css)
+
                     with ui.row():
                         switch_tongyi_history_enable = ui.switch('上下文记忆', value=config.get("tongyi", "history_enable")).style(switch_internal_css)
                         input_tongyi_history_max_len = ui.input(label='最大记忆长度', value=config.get("tongyi", "history_max_len"), placeholder='最长能记忆的问答字符串长度，超长会丢弃最早记忆的内容，请慎用！配置过大可能会有丢大米')