change, test

app.py

@@ -7,17 +7,16 @@ model_id = "TheBloke/KafkaLM-70B-German-V0.1-GGUF"
 model_filename = "kafkalm-70b-german-v0.1.Q5_K_M.gguf"
 model_path = hf_hub_download(repo_id=model_id, filename=model_filename)
 
+def generate_text(user_input, system_prompt):
 
+    # Initialize the Llama model
+    llm = Llama(
+        model_path=model_path,  # Use the downloaded model file
+        n_ctx=4096,             # Adjust based on the model's max sequence length
+        n_threads=8,            # Tailor to your system
+        n_gpu_layers=35         # Set based on your GPU's capability
+    )
 
-# Initialize the Llama model
-llm = Llama(
-    model_path=model_path,  # Use the downloaded model file
-    n_ctx=4096,             # Adjust based on the model's max sequence length
-    n_threads=8,            # Tailor to your system
-    n_gpu_layers=35         # Set based on your GPU's capability
-)
-
-def generate_text(user_input, system_prompt):
     # Combine the system and user prompts
     prompt = f"\n{system_prompt.strip()}</s>\n\n{user_input.strip()}</s>\n"