Update r-story-test.py

r-story-test.py

@@ -78,7 +78,6 @@ print("\n최종 문장 수:", len(texts))
 
 # In[6]:
 
-
 from langchain.vectorstores import FAISS
 from langchain.embeddings import HuggingFaceEmbeddings
 
@@ -118,7 +117,7 @@ from transformers import pipeline
 generator = pipeline(
     "text-generation",
     model="kakaocorp/kanana-nano-2.1b-instruct",
-    device=0  
+    device= -1
 )
 
 
@@ -139,9 +138,10 @@ vectorstore = FAISS.load_local("solo_leveling_faiss_ko", embedding_model, allow_
 
 model_name = "kakaocorp/kanana-nano-2.1b-instruct"
 tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16).to("cuda")
+model = AutoModelForCausalLM.from_pretrained(
+    model_name, torch_dtype=torch.float32).to("cpu")
 
-llm_pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer, max_new_tokens=256)
+llm_pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer, max_new_tokens=128)
 llm = HuggingFacePipeline(pipeline=llm_pipeline)
 
 custom_prompt = PromptTemplate(