Update ultravox_processing.py

ultravox_processing.py

@@ -171,14 +171,14 @@ class UltravoxProcessor(transformers.ProcessorMixin):
 
                     start_idx = len(
                         self.tokenizer.encode(
-                            t[: t.index(self.audio_placeholder)],
+                            t.split(self.audio_placeholder)[0],
                             add_special_tokens=False,
                         )
                     )
                     data["audio_token_start_idx"].append(start_idx)
 
                     # Replace the audio placeholder with the audio token.
-                    #   e.g. "Transcribe\n<|audio|>" -> "Transcribe </s></s></s></s></s></s></s></s>"
+                    #   e.g. "Transcribe <|audio|>" -> "Transcribe </s></s></s></s></s></s></s></s>"
                     #        where the number of </s> is the number of audio frames.
                     t = t.replace(
                         self.audio_placeholder,