Update app.py

app.py

@@ -1,12 +1,36 @@
+import os
 import urllib
 import chatglm_cpp
 import gradio as gr
 
-DEFAULT_MODEL_PATH = "chatglm3-6b.bin"
+import requests
 
-pipeline = chatglm_cpp.Pipeline(DEFAULT_MODEL_PATH)
-system_message = chatglm_cpp.ChatMessage(role="system", content="请你现在扮演一个软件工程师，名字叫做贺英旭。你需要以这个身份和朋友们对话。")
+endpoint_url = os.getenv('ENDPOINT_URL')
+personal_secret_token = os.getenv('PERSONAL_HF_TOKEN')
 
+turn_breaker = os.getenv('TURN_BREAKER')
+system_symbol = os.getenv('SYSTEM_SYMBOL')
+user_symbol = os.getenv('USER_SYMBOL')
+assistant_symbol = os.getenv('ASSISTANT_SYMBOL')
+
+headers = {
+	"Accept" : "application/json",
+    "Authorization": f"Bearer {personal_secret_token}",
+	"Content-Type": "application/json" 
+}
+
+def query(payload):
+	# response = requests.post(endpoint_url, headers=headers, json=payload)
+	# return response.json()
+    return payload['inputs']
+
+# output = query({
+# 	"inputs": "你啥比",
+# 	"parameters": {
+# 		"max_new_tokens": 150
+# 	}
+# })
+# system_message = chatglm_cpp.ChatMessage(role="system", content="请你现在扮演一个软件工程师，名字叫做贺英旭。你需要以这个身份和朋友们对话。")
 
 def respond(
     message,
@@ -16,30 +40,36 @@ def respond(
     temperature,
     top_p,
 ):
-    messages = [chatglm_cpp.ChatMessage(role="system", content=system_message)]
+    # messages = [chatglm_cpp.ChatMessage(role="system", content=system_message)]
     # messages = [{"role": "system", "content": system_message}]
+    all_messages = [system_message]
 
     for val in history:
         if val[0]:
             # messages.append({"role": "user", "content": val[0]})
-            messages.append(chatglm_cpp.ChatMessage(role="user", content=val[0]))
+            all_messages.append(user_symbol+val[0])
+            # messages.append(chatglm_cpp.ChatMessage(role="user", content=val[0]))
         if val[1]:
-            messages.append(chatglm_cpp.ChatMessage(role="assistant", content=val[1]))
+            all_messages.append(assistant_symbol+val[1])
+            # messages.append(chatglm_cpp.ChatMessage(role="assistant", content=val[1]))
 
-    messages.append(chatglm_cpp.ChatMessage(role="user", content=message))
+    # messages.append(chatglm_cpp.ChatMessage(role="user", content=message))
+    all_messages.append(user_symbol+message)
 
     generation_kwargs = dict(
         max_new_tokens=max_new_tokens,
         do_sample=temperature > 0,
         top_p=top_p,
         temperature=temperature,
-        stream=True,
+        # stream=True,
     )
 
-    response = ""
-    for chunk in pipeline.chat(messages, **generation_kwargs):
-        response += chunk.content
-        yield response
+    response = query({
+        "inputs": turn_breaker.join(all_messages),
+        "parameters": generation_kwargs
+    })
+
+    return response
 
 
 """
@@ -48,7 +78,7 @@ For information on how to customize the ChatInterface, peruse the gradio docs: h
 demo = gr.ChatInterface(
     respond,
     additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
+        gr.Textbox(value="请你现在扮演一个软件工程师，名字叫做贺英旭。你需要以这个身份和朋友们对话。", label="System message"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
         gr.Slider(