Updated to new version llama-cpp

requirements.txt

@@ -3,4 +3,4 @@ requests>=2.31.0
 pathlib2>=2.3.7
 pinecone>=3.0.0
 numpy>=1.21.0
-llama-cpp-python==0.1.9
+llama-cpp-python==0.3.16

scripts/llm_service.py

@@ -199,12 +199,12 @@ Always be helpful, accurate, and focused on the user's specific needs."""
                 # )
                 self.llm = Llama(
                     model_path=self.model_path,
-                    # cache_dir=Path('models'),
+                    cache_dir=Path('models'),
                     seed=42,
                     n_ctx=self.n_ctx,
-                    # verbose=False,
-                    # n_gpu_layers=self.n_gpu_layers,
-                    # n_threads=self.n_threads,
+                    verbose=False,
+                    n_gpu_layers=self.n_gpu_layers,
+                    n_threads=self.n_threads,
                 )
                 
                 self.is_loaded = True