Update app.py

app.py

@@ -1,24 +1,38 @@
+# build on your original chatbot from the previous lesson
+# a basic chatbot from the previous lesson is below -- edit it to incorporate the changes described above
+
 import gradio as gr
-from huggingface_hub import InferenceClient
+from huggingface_hub import InferenceClient #imports huggingface models
 
 client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
 def respond(message, history):
-    
-    messages = [{"role": "system", "content": "You are a friendly chatbot."}]
-    
+    messages = [{"role": "system", "content": "I am a kind chatbot."}]
+
+# 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})
-    
-    response = client.chat_completion(
+
+    # 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=100
-    )
-    
-    return response['choices'][0]['message']['content'].strip()
+        max_tokens=100,
+        temperature=.1,
+        stream=True):
+
+        # add the tokens to the output content
+          token = message.choices[0].delta.content # capture the most recent toke
+          response += token # Add it to the response
+          yield response # yield the response:
 
-chatbot = gr.ChatInterface(respond, type="messages")
+chatbot = gr.ChatInterface(respond)
 
 chatbot.launch()