Debug using Gemini

app.py

@@ -1,40 +1,47 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
+from sentence_transformers import SentenceTransformer
+import torch
+import numpy as np
 
 # this client will handle making requests to the model to generate responses
 client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
 def respond(message, history):
 
-    system_message = "You are a friendly tour guide chatbot. You always respond with an upbeat attitude!"
-    
+    system_message = "You are a kitchen sous chef. You always respond with a knowledgeable and upbeat attitude!"
+
     # initialize a list of dictionaries to store the messages
-    messages = [{"role": "system", 
+    messages = [{"role": "system",
                  "content": system_message}]
-    
-    # add all previous messages to the messages list 
+
+    # add all previous messages to the messages list
     if history:
         messages.extend(history)
 
     # add the current user’s message to the messages list
     messages.append({"role": "user", "content": message})
 
-    # makes the chat completion API call, 
+    # makes the chat completion API call,
     # sending the messages and other parameters to the model
     # implements streaming, where one word/token appears at a time
     response = ""
 
     # iterate through each message in the method
-    for message in client.chat_completion(
-        messages,
-        max_tokens=200,
-        temperature=0.9,
-        stream=True
-    ):
-        # add the tokens to the output content
-        token = message.choices[0].delta.content # capture the most recent token
-        response += token # Add it to the response
-        yield response # yield the response
+    try:
+      for message in client.chat_completion(
+          messages,
+          max_tokens=200,
+          temperature=0.9,
+          stream=True
+      ):
+          # add the tokens to the output content
+          token = message.choices[0].delta.content # capture the most recent token
+          response += token # Add it to the response
+          yield response # yield the response
+
+    except Exception as e:
+        print(f"An error occurred: {e}")
 
 # Create the Gradio chat interface (using Gradio's ChatInterface)
 title = "# SousChef AI 🍳"
@@ -55,14 +62,14 @@ with gr.Blocks() as chatbot:
         with gr.Column(scale=1):
             gr.Markdown(title)
             gr.Markdown(topics)
-        
+
         with gr.Column(scale=2):
-            gr.ChatInterface(respond, 
-                             type="messages", 
+            gr.ChatInterface(respond,
+                             type="messages",
                              examples = ["How do I know if my chicken is fully cooked?", "What toppings should I add to my ramen?", "What's the best recipe for homemade pizza dough?"],
                              theme='d8ahazard/material_design_rd'
                             )
     with gr.Row():
         gr.Markdown(disclaimer)
 
-chatbot.launch(debug=True)
+chatbot.launch(debug=False)