diff --git a/egs/wenetspeech4tts/TTS/local/compute_neural_codec_and_prepare_text_tokens.py b/egs/wenetspeech4tts/TTS/local/compute_neural_codec_and_prepare_text_tokens.py
index 5494bf3400..4e0a47c68d 100755
--- a/egs/wenetspeech4tts/TTS/local/compute_neural_codec_and_prepare_text_tokens.py
+++ b/egs/wenetspeech4tts/TTS/local/compute_neural_codec_and_prepare_text_tokens.py
@@ -516,9 +516,15 @@ def main():
             for idx, part in enumerate(cut_sets):
                 if args.audio_extractor:
                     if args.audio_extractor == "Encodec":
-                        storage_path = f"{args.output_dir}/{args.prefix}_encodec_{partition}_{idx if split > 1 else ''}"
+                        if split > 1:
+                            storage_path = f"{args.output_dir}/{args.prefix}_encodec_{partition}_{idx}"
+                        else:
+                            storage_path = f"{args.output_dir}/{args.prefix}_encodec_{partition}"
                     else:
-                        storage_path = f"{args.output_dir}/{args.prefix}_fbank_{partition}_{idx if split > 1 else ''}"
+                        if split > 1:
+                            storage_path = f"{args.output_dir}/{args.prefix}_fbank_{partition}_{idx}"
+                        else:
+                            storage_path = f"{args.output_dir}/{args.prefix}_fbank_{partition}"
 
                     if args.prefix.lower() in [
                         "ljspeech",
@@ -587,9 +593,11 @@ def main():
                         ].normalized_text, "normalized_text is None"
 
                 # Save each part with an index if split > 1
-                cuts_filename = (
-                    f"{prefix}cuts_{partition}.{idx if split > 1 else ''}.{args.suffix}"
-                )
+                if split > 1:
+                    cuts_filename = f"{prefix}cuts_{partition}.{idx}.{args.suffix}"
+                else:
+                    cuts_filename = f"{prefix}cuts_{partition}.{args.suffix}"
+
                 part.to_file(f"{args.output_dir}/{cuts_filename}")
                 logging.info(f"Saved {cuts_filename}")
 
diff --git a/egs/wenetspeech4tts/TTS/valle/requirements.txt b/egs/wenetspeech4tts/TTS/valle/requirements.txt
new file mode 100644
index 0000000000..06958dbeaf
--- /dev/null
+++ b/egs/wenetspeech4tts/TTS/valle/requirements.txt
@@ -0,0 +1,2 @@
+phonemizer==3.2.1
+git+https://github.com/facebookresearch/encodec.git
\ No newline at end of file