Update app.py

app.py

@@ -27,7 +27,7 @@ if not torch.cuda.is_available():
 
 
 if torch.cuda.is_available():
-    model_id = "wangzhang/ChatSDB-hf"
+    model_id = "wangzhang/ChatSDB-tb-testing"
     model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16, device_map="auto")
     tokenizer = AutoTokenizer.from_pretrained(model_id)
     tokenizer.use_default_system_prompt = False
@@ -49,9 +49,9 @@ def generate(
         conversation.append({"role": "system", "content": system_prompt})
     for user, assistant in chat_history:
         conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
-    conversation.append({"role": "user", "content": "### Instruction: \n" + "根据巨杉数据库SequoiaDB的相关问题进行回答。\n" + "### Input:\n" + message + "\n ### Response:"})
+    conversation.append({"role": "user", "content": message)
 
-    chat = tokenizer.apply_chat_template([{"role": "user", "content": "### Instruction: \n" + "根据巨杉数据库SequoiaDB的相关问题进行回答。\n" + "### Input:\n" + message + "\n ### Response:"}], tokenize=False)
+    chat = tokenizer.apply_chat_template(conversation, tokenize=False)
     inputs = tokenizer(chat, return_tensors="pt", add_special_tokens=False).to("cuda")
     if len(inputs) > MAX_INPUT_TOKEN_LENGTH:
         inputs = inputs[-MAX_INPUT_TOKEN_LENGTH:]