Update llm_generator.py

llm_generator.py

@@ -26,17 +26,16 @@ config = {'max_new_tokens': 256,
           'context_length': 4096, # Set to max for Chat Summary, Llama-2 has a max context length of 4096
           }
 
-# Load model directly
-from transformers import AutoModel
-model = AutoModel.from_pretrained("TheBloke/Llama-2-7B-Chat-GGUF")
 
-llm = CTransformers(model,
+
+llm = CTransformers(model=os.path.join(script_dir, 'transformers', 'llama-2-7b-chat.Q2_K.gguf'),
                     callbacks=[StreamingStdOutCallbackHandler()],
                     config=config)
 
 
 
 
+
 # Define system and user message templates
 with open(os.path.join(script_dir, 'prompts', 'system_message_template.txt'), 'r') as file:
     system_message_template = file.read().replace('\n', '')