commit 000435

app.py

@@ -79,9 +79,12 @@ chain = RunnableBranch(
 
 # Reply generator
 @GPU
-def generate_reply(message):
-    result = chain.invoke({"question": message})
-    return result.strip()
+def generate_full_reply(message, history):
+    system_prompt = "You are a helpful and concise AI assistant. Answer briefly and politely, in one or two sentences."
+    messages = [{"role": "system", "content": system_prompt}] + history + [{"role": "user", "content": message}]
+    prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+    output = pipe(prompt, max_new_tokens=512, do_sample=True, temperature=0.7)[0]['generated_text']
+    return output.split(prompt)[-1].strip()
 
 # Flask routes
 @app.route("/")