Update app.py

app.py

@@ -13,9 +13,8 @@ from huggingface_hub import hf_hub_download
 llm = None
 llm_model = None
 
-# 모델 이름과 경로를 정의 (전역 변수로 활용)
+# 모델 이름과 경로를 정의
 MISTRAL_MODEL_NAME = "Private-BitSix-Mistral-Small-3.1-24B-Instruct-2503.gguf"
-LLAMA_MODEL_NAME = "Meta-Llama-3-70B-Instruct-Q3_K_M.gguf"
 
 # 모델 다운로드
 model_path = hf_hub_download(
@@ -78,11 +77,7 @@ css = """
 """
 
 def get_messages_formatter_type(model_name):
-    if "Llama" in model_name:
-        return MessagesFormatterType.LLAMA_3
-    elif "unsloth" in model_name:
-        return MessagesFormatterType.CHATML
-    elif "Mistral" in model_name or "BitSix" in model_name:
+    if "Mistral" in model_name or "BitSix" in model_name:
         return MessagesFormatterType.CHATML  # Mistral 계열 모델은 ChatML 형식 사용
     else:
         raise ValueError(f"Unsupported model: {model_name}")
@@ -91,7 +86,6 @@ def get_messages_formatter_type(model_name):
 def respond(
     message,
     history: list[tuple[str, str]],
-    model_choice,
     system_message,
     max_tokens,
     temperature,
@@ -102,22 +96,18 @@ def respond(
     global llm
     global llm_model
     
-    chat_template = get_messages_formatter_type(model_choice)
+    chat_template = get_messages_formatter_type(MISTRAL_MODEL_NAME)
     
     # 모델 파일 경로 확인
-    if model_choice == MISTRAL_MODEL_NAME:
-        model_path = os.path.join("./models", MISTRAL_MODEL_NAME)
-    else:
-        model_path = os.path.join("./models", model_choice)
+    model_path = os.path.join("./models", MISTRAL_MODEL_NAME)
     
-    print(f"Selected model: {model_choice}")
     print(f"Model path: {model_path}")
     
     if not os.path.exists(model_path):
         print(f"Warning: Model file not found at {model_path}")
         print(f"Available files in ./models: {os.listdir('./models')}")
     
-    if llm is None or llm_model != model_choice:
+    if llm is None or llm_model != MISTRAL_MODEL_NAME:
         llm = Llama(
             model_path=model_path,
             flash_attn=True,
@@ -125,7 +115,7 @@ def respond(
             n_batch=1024,
             n_ctx=8192,
         )
-        llm_model = model_choice
+        llm_model = MISTRAL_MODEL_NAME
     
     provider = LlamaCppPythonProvider(llm)
 
@@ -232,13 +222,7 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="violet", secondary_hue="violet"
                 )
                 submit_btn = gr.Button("전송", variant="primary")
         
-        with gr.Column(scale=1):
-            model_choice = gr.Dropdown(
-                [MISTRAL_MODEL_NAME, LLAMA_MODEL_NAME],
-                value=MISTRAL_MODEL_NAME,
-                label="모델"
-            )
-            
+        with gr.Column(scale=1):            
             system_message = gr.Textbox(
                 value="You are a deep thinking AI, you may use extremely long chains of thought to deeply consider the problem and deliberate with yourself via systematic reasoning processes to help come to a correct solution prior to answering. You should enclose your thoughts and internal monologue inside tags, and then provide your solution or response to the problem.",
                 label="시스템 메시지",
@@ -253,12 +237,11 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="violet", secondary_hue="violet"
             
             with gr.Row():
                 clear_btn = gr.Button("대화 초기화")
-                retry_btn = gr.Button("다시 시도")
-    
+                
     # Event handlers
     submit_btn.click(
         fn=respond,
-        inputs=[msg, chatbot, model_choice, system_message, max_tokens, temperature, top_p, top_k, repeat_penalty],
+        inputs=[msg, chatbot, system_message, max_tokens, temperature, top_p, top_k, repeat_penalty],
         outputs=[chatbot],
         queue=True
     ).then(
@@ -269,7 +252,7 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="violet", secondary_hue="violet"
     
     msg.submit(
         fn=respond,
-        inputs=[msg, chatbot, model_choice, system_message, max_tokens, temperature, top_p, top_k, repeat_penalty],
+        inputs=[msg, chatbot, system_message, max_tokens, temperature, top_p, top_k, repeat_penalty],
         outputs=[chatbot],
         queue=True
     ).then(
@@ -279,18 +262,6 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="violet", secondary_hue="violet"
     )
     
     clear_btn.click(lambda: None, None, chatbot, queue=False)
-    
-    retry_btn.click(
-        fn=lambda x: x[:-1],
-        inputs=[chatbot],
-        outputs=[chatbot],
-        queue=False
-    ).then(
-        fn=respond,
-        inputs=[chatbot[-1][0] if len(chatbot) > 0 else "", chatbot, model_choice, system_message, max_tokens, temperature, top_p, top_k, repeat_penalty],
-        outputs=[chatbot],
-        queue=True
-    )
 
 if __name__ == "__main__":
     demo.launch()