Update app.py

app.py

@@ -32,13 +32,13 @@ scheduler = CommitScheduler(
 
 pc = Pinecone(api_key=os.environ.get("PINECONE"))
 index = pc.Index("commonsense")
-
-device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+"""
+device = torch.device("cuda" if torch.cuda.is_available() else "CPU")
 retriever_tokenizer = AutoTokenizer.from_pretrained("psyche/dpr-longformer-ko-4096")
 retriever = AutoModel.from_pretrained("psyche/dpr-longformer-ko-4096")
 retriever.eval()
 retriever.to(device)
-
+"""
 def save_json(question: str, answer: str) -> None:
     with scheduler.lock:
         with JSON_DATASET_PATH.open("a") as f:
@@ -82,7 +82,7 @@ def generate(
     for user, assistant in chat_history:
         conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
 
-    
+    """    
     retriever_inputs = retriever_tokenizer([message], max_length=1024, truncation=True, return_tensors="pt")
     retriever_inputs = {k:v.to(retriever.device) for k,v in retriever_inputs.items()}
     with torch.no_grad():
@@ -99,7 +99,7 @@ def generate(
     results = [result for result in results["matches"] if result["score"] > 0.6]
     if len(results) > 0:
         message = results[0]["metadata"]["text"] + f"\n\n위 문맥을 참고하여 질문 '{message}'에 답하면?"
-        
+    """
     conversation.append({"role": "user", "content": message })
     
     input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt", add_generation_prompt=True)