Update app.py

app.py

@@ -9,13 +9,10 @@ For more information on `huggingface_hub` Inference API support, please check th
 client = InferenceClient("kronos25/Temporal_Chatbot")
 
 def predict(message, history):
-  model = T5ForConditionalGeneration.from_pretrained('')
-  tokenizer = T5Tokenizer.from_pretrained('')
-  input = message + '\n'
-  inputs = tokenizer(input, return_tensors="pt")
-  outputs = model.generate(**inputs,max_length=100)
-  model_result = tokenizer.decode(outputs[0], skip_special_tokens=True)
-  return model_result + '\n'
+  for message in client.chat_completion(messages,max_tokens=max_tokens):
+    token = message.choices[0].delta.content
+    response += token
+    yield response
 
 gr.ChatInterface(
     predict,