Update app.py

app.py

@@ -5,6 +5,7 @@ import nltk
 import copy
 import time
 import vllm
+import torch
 import spacy
 import shutil
 import msgpack
@@ -169,7 +170,7 @@ def load_models_and_documents():
             quantization="bitsandbytes",
             load_format="bitsandbytes",
             gpu_memory_utilization=0.9,
-            dtype='float16'
+            dtype=torch.float16
         )
         model = models.VLLM(llm)