diff --git a/egs/librispeech/ASR/conformer_ctc/pretrained.py b/egs/librispeech/ASR/conformer_ctc/pretrained.py
index b4142ab89..eee8b94d4 100755
--- a/egs/librispeech/ASR/conformer_ctc/pretrained.py
+++ b/egs/librispeech/ASR/conformer_ctc/pretrained.py
@@ -20,7 +20,7 @@
 import argparse
 import logging
 import math
-import os
+from pathlib import Path
 from typing import List
 
 import k2
@@ -298,8 +298,8 @@ def main():
 
     if params.method == "ctc-decoding":
         logging.info("Use CTC decoding")
-        if not os.path.exists(params.bpe_model):
-            raise ValueError("The path to bpe.model is required!")
+        if not Path(params.bpe_model).exists():
+            raise ValueError(f"The path to {params.bpe_model} doesn't exist!")
 
         bpe_model = spm.SentencePieceProcessor()
         bpe_model.load(params.bpe_model)
@@ -333,10 +333,10 @@ def main():
         "whole-lattice-rescoring",
         "attention-decoder",
     ]:
-        if not os.path.exists(params.HLG):
-            raise ValueError("The path to HLG.pt is required!")
-        if not os.path.exists(params.words_file):
-            raise ValueError("The path to words.txt is required!")
+        if not Path(params.HLG).exists():
+            raise ValueError(f"The path to {params.HLG} doesn't exist!")
+        if not Path(params.words_file).exists():
+            raise ValueError(f"The path to {params.words_file} doesn't exist!")
 
         logging.info(f"Loading HLG from {params.HLG}")
         HLG = k2.Fsa.from_dict(torch.load(params.HLG, map_location="cpu"))