Fix exception when generating the subtitles for an audio with no speech

[jordimas]

When generating the subtitles for an audio with no speech, an exception is thrown because the code assumes that there will be always a segment at index 0 which is not the case.

  File "/lib/python3.11/site-packages/whisper/transcribe.py", line 453, in cli
    writer(result, audio_path, writer_args)
  File "/lib/python3.11/site-packages/whisper/utils.py", line 255, in write_all
    writer(result, file, options)
  File "/lib/python3.11/site-packages/whisper/utils.py", line 85, in __call__
    self.write_result(result, file=f, options=options)
  File "/lib/python3.11/site-packages/whisper/utils.py", line 195, in write_result
    for start, end, text in self.iterate_result(result, options):
  File "/lib/python3.11/site-packages/whisper/utils.py", line 148, in iterate_result
    if "words" in result["segments"][0]:
                  ~~~~~~~~~~~~~~~~~~^^^
IndexError: list index out of range

[gmat]

this change should be approved it fixes crashes on srt and vtt outputs when no speech identified in audio. Tested with last release and python 3.9. Will works with all python versions