Update app.py

app.py

@@ -1,4 +1,3 @@
-!pip install gradio
 import gradio as gr
 from transformers import GPT2LMHeadModel, GPT2Tokenizer
 
@@ -6,8 +5,8 @@ model_name = "anezatra/gpt2_openassistant_guanaco"
 model = GPT2LMHeadModel.from_pretrained(model_name)
 tokenizer = GPT2Tokenizer.from_pretrained(model_name)
 
-def respond(message):
-    prompt = f"{message}"
+def respond(message, history):
+    prompt = "\n".join(history + [message])
 
     inputs = tokenizer.encode(prompt, return_tensors="pt")
     outputs = model.generate(
@@ -20,7 +19,10 @@ def respond(message):
         do_sample=True,
     )
 
-    response = tokenizer.decode(outputs[0], skip_special_tokens=True).split("### Assistant:")[-1].strip()
+    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    response = response.split("### Assistant:")[-1].strip()
+    if "### Human:" in response:
+        response = response.split("### Human:")[1].strip()
     return response
 
 banner = gr.HTML("""
@@ -40,11 +42,11 @@ with gr.Blocks(theme=gr.Theme.from_hub('gradio/monochrome')) as demo:
         clear = gr.Button("Clear chat")
         submit = gr.Button("Send message")
 
-    def user_input(user_message):
-        response = respond(user_message)
-        return "", [[user_message, response]]
+    def user_input(user_message, history):
+        response = respond(user_message, history)
+        return "", history + [[user_message, response]]
 
-    msg.submit(user_input, [msg], [msg, chatbot], queue=False)
+    msg.submit(user_input, [msg, chatbot], [msg, chatbot], queue=False)
     clear.click(lambda: None, None, chatbot, queue=False)
     submit.click(lambda: msg.submit(), None, chatbot, queue=False)