Update src/tinny_llama.py

src/tinny_llama.py

@@ -132,8 +132,7 @@ def initialize_llm() -> Optional[HuggingFacePipeline]:
             do_sample=True,
             temperature=CONFIG["temperature"],
             top_p=CONFIG["top_p"],
-            truncation=True,
-            device=-1  # Explicitly set to CPU
+            truncation=True
         )
         logger.info("LLM initialized successfully.")
         return HuggingFacePipeline(pipeline=text_gen)