init inference

app.py

@@ -17,6 +17,8 @@ lora_adapter_path = hf_hub_download(repo_id=adapter_repo, filename="llama_lora_a
 
 # Define the llama-cli path explicitly
 llama_cli_path = "./llama.cpp/build/bin/llama-cli"
+if not os.access(llama_cli_path, os.X_OK):  # Check if the file is executable
+    os.chmod(llama_cli_path, 0o755)  # Set executable permissions
 
 # Function to run `llama-cli` with base model and adapter
 def run_llama_cli(prompt):