Update app.py

app.py

@@ -75,16 +75,20 @@ def format_prompt(user_prompt, chat_history):
     return prompt
 
 # Model inference function
-def model_inference(user_prompt, web_search):
+def model_inference(prompt, web_search):
+    for response in fetch_response(prompt, web_search):
+        yield response
+
+def fetch_response(prompt, web_search):
     if web_search:
         # Perform web search and generate text based on the retrieved results
-        web_results = search(user_prompt)
+        web_results = search(prompt)
         web2 = ' '.join([f"Link: {res['link']}\nText: {res['text']}\n\n" for res in web_results])
         client = InferenceClient("mistralai/Mistral-7B-Instruct-v0.3")
         generate_kwargs = dict(max_new_tokens=4000, do_sample=True)
         formatted_prompt = format_prompt(
-            f"""You are OpenGPT 4o... [USER] {user_prompt} [WEB] {web2} [OpenGPT 4o]""",
-            [(user_prompt, web2)])
+            f"""You are OpenGPT 4o... [USER] {prompt} [WEB] {web2} [OpenGPT 4o]""",
+            [(prompt, web2)])
         stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
         output = ""
         for response in stream:
@@ -95,7 +99,7 @@ def model_inference(user_prompt, web_search):
         # Use the microsoft/Phi-3-mini-4k-instruct model for generating text based on user prompts
         client = InferenceClient("microsoft/Phi-3-mini-4k-instruct")
         generate_kwargs = dict(max_new_tokens=5000, do_sample=True)
-        formatted_prompt = format_prompt(f"""You are OpenGPT 4o... [USER] {user_prompt} [OpenGPT 4o]""", [(user_prompt, )])
+        formatted_prompt = format_prompt(f"""You are OpenGPT 4o... [USER] {prompt} [OpenGPT 4o]""", [(prompt, )])
         stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
         output = ""
         for response in stream:
@@ -104,20 +108,20 @@ def model_inference(user_prompt, web_search):
             yield output
 
 # Create a chatbot interface with a Fetch button
-def fetch_response(prompt, web_search):
+def start_inference(prompt, web_search):
     for response in model_inference(prompt, web_search):
         return response
 
 chatbot = gr.Interface(
-    fn=fetch_response,
+    fn=start_inference,
     inputs=[
-        gr.Textbox(label="User Prompt"),
+        gr.Textbox(label="User Prompt", placeholder="Enter your prompt here..."),
         gr.Checkbox(label="Enable Web Search", value=False),
         "button"
     ],
-    outputs=gr.Textbox(label="Response"),
+    outputs=gr.Textbox(label="Response", placeholder="Responses will appear here..."),
     live=True
 )
 
 # Launch the Gradio interface
-chatbot.launch()
+chatbot.launch(share=True)