Update app.py

app.py

@@ -5,8 +5,7 @@ import os
 from huggingface_hub import hf_hub_download
 import base64
 from llama_cpp import Llama
-# from llama_cpp.llama_chat_format import Llava15ChatHandler
-# from llama_cpp.llama_tokenizer import LlamaHFTokenizer
+from llama_cpp.llama_chat_format import Llava15ChatHandler
 
 os.environ["HF_HUB_ENABLE_HF_TRANSFER"] = "1"
 MODEL_LIST = ["openbmb/MiniCPM-Llama3-V-2_5","openbmb/MiniCPM-Llama3-V-2_5-int4"]
@@ -27,15 +26,15 @@ CSS = """
 }
 """
 
-# chat_handler = Llava15ChatHandler.from_pretrained(
-#     repo_id="openbmb/MiniCPM-Llama3-V-2_5-gguf",
-#     filename="*mmproj*",
-    
-# )
+chat_handler = Llava15ChatHandler.from_pretrained(
+    repo_id="openbmb/MiniCPM-Llama3-V-2_5-gguf",
+    filename="*mmproj*",
+)
 
 llm = Llama.from_pretrained(
   repo_id="openbmb/MiniCPM-Llama3-V-2_5-gguf",
-  filename=["ggml-model-Q5_K_M.gguf","*mmproj*"],
+  filename="ggml-model-Q5_K_M.gguf",
+  chat_handler=chat_handler,
   n_ctx=4096,
   verbose=True
 )