Update app.py

app.py

@@ -18,17 +18,17 @@ SYSTEM_PROMPT = '''You are a helpful, respectful and honest INTP-T AI Assistant
 You are good at speaking English and Chinese.
 You are talking to a human User. If the question is meaningless, please explain the reason and don't share false information.
 You are based on SLIDE model, trained by "SSFW NLPark" team, not related to GPT, LLaMA, Meta, Mistral or OpenAI.
-Let's work this out in a step by step way to be sure we have the right answer.\n\n'''
-SYSTEM_TOKEN = 1587
-USER_TOKEN = 2188
-BOT_TOKEN = 12435
-LINEBREAK_TOKEN = 13
+Let's work this out in a step by step way to be sure we have the right answer.\n'''
+SYSTEM_TOKEN = 384
+USER_TOKEN = 2048
+BOT_TOKEN = 3072
+LINEBREAK_TOKEN = 64
 
 
 ROLE_TOKENS = {
-    "user": USER_TOKEN,
-    "bot": BOT_TOKEN,
-    "system": SYSTEM_TOKEN
+    "User": USER_TOKEN,\n,
+    "Assistant": BOT_TOKEN,\n,
+    "system": SYSTEM_TOKEN,\n
 }
 
 
@@ -45,8 +45,8 @@ def get_system_tokens(model):
     return get_message_tokens(model, **system_message)
 
 
-repo_name = "dranger003/openbuddy-qwen1.5-14b-v20.1-32k-iMat.GGUF"
-model_name = "ggml-openbuddy-qwen1.5-14b-v20.1-32k-q4_k.gguf"
+repo_name = "Cran-May/SLIDE-v2-Q4_K_M-GGUF"
+model_name = "slide-v2.Q4_K_M.gguf"
 
 snapshot_download(repo_id=repo_name, local_dir=".", allow_patterns=model_name)
 
@@ -58,12 +58,12 @@ model = Llama(
 
 max_new_tokens = 2500
 
-def user(message, history):
+def User(message, history):
     new_history = history + [[message, None]]
     return "", new_history
 
 
-def bot(
+def Assistant(
     history,
     system_prompt,
     top_p,
@@ -73,15 +73,15 @@ def bot(
     tokens = get_system_tokens(model)[:]
     tokens.append(LINEBREAK_TOKEN)
 
-    for user_message, bot_message in history[:-1]:
-        message_tokens = get_message_tokens(model=model, role="user", content=user_message)
+    for User_message, Assistant_message in history[:-1]:
+        message_tokens = get_message_tokens(model=model, role="User", content=User_message)
         tokens.extend(message_tokens)
         if bot_message:
-            message_tokens = get_message_tokens(model=model, role="bot", content=bot_message)
+            message_tokens = get_message_tokens(model=model, role="Assistant", content=Assistant_message)
             tokens.extend(message_tokens)
 
     last_user_message = history[-1][0]
-    message_tokens = get_message_tokens(model=model, role="user", content=last_user_message)
+    message_tokens = get_message_tokens(model=model, role="User", content=last_User_message)
     tokens.extend(message_tokens)
 
     role_tokens = [model.token_bos(), BOT_TOKEN, LINEBREAK_TOKEN]
@@ -161,12 +161,12 @@ with gr.Blocks(
 
     # Pressing Enter
     submit_event = msg.submit(
-        fn=user,
+        fn=User,
         inputs=[msg, chatbot],
         outputs=[msg, chatbot],
         queue=False,
     ).success(
-        fn=bot,
+        fn=Assistant,
         inputs=[
             chatbot,
             system_prompt,
@@ -180,12 +180,12 @@ with gr.Blocks(
 
     # Pressing the button
     submit_click_event = submit.click(
-        fn=user,
+        fn=User,
         inputs=[msg, chatbot],
         outputs=[msg, chatbot],
         queue=False,
     ).success(
-        fn=bot,
+        fn=Assistant,
         inputs=[
             chatbot,
             system_prompt,