Update app.py

app.py

@@ -14,11 +14,20 @@ examples = [
 ]
 
 def process_stream(instruction, temperature, top_p, top_k, max_new_tokens, seed):
-    streamer = chatglm_llm.generate(prompt=instruction,
+    if "[SEP]" not in instruction:
+        streamer = chatglm_llm.generate(prompt=instruction,
         temperature=temperature,
         top_p=top_p,top_k=top_k,max_length=max_new_tokens,
         stream = True
     )
+    else:
+        history = instruction.split("[SEP]")
+        chatglm_llm.chat(
+            history=history,
+            temperature=temperature,
+            top_p=top_p,top_k=top_k,max_length=max_new_tokens,
+            , do_sample=False
+        )
     response = ""
     for new_text in streamer:
         response += new_text
@@ -31,7 +40,6 @@ with gr.Blocks(
 ) as demo:
     gr.Markdown(
         """<h1><center> ChatGLM2 on CPU in CPP 🐒</center></h1>
-
         This demo uses the [chatglm.cpp](https://github.com/li-plus/chatglm.cpp) library on 2 CPU cores.
         """
     )