On my Raspberry Pi 3 Model B+, with Raspbian Lite, I installed Rhasspy 2.5.0-pre in a docker.
When I change ‘Speech to Text’ to Pocketsphinx, I got the message: ‘Rhasspy needs to download some files for your profile (9 missing, 379.3 MiB)’
When I click on de Downloadbutton, there is the following error:
‘Part size mismatch (got 170747 byte(s), expected 13859764)’)
When I click again then sometimes for other parts of the same file (part-00, part-01, part-02 and part-03).
What is wrong here?

I think I’m having a similar issue. Did you solve it?
After updating to 2.5.10 I had to download some Files:

  • gruut/de-de.tar.gz
  • larynx/hifi_gan-universal_large.tar.gz

But I’m getting the following issue as a popup:

DownloadFailedException: ('', 'File size mismatch (got 20218377 byte(s), expected 20219379)')

And this is my log output if it can help someone:

Traceback (most recent call last):
  File "/usr/lib/rhasspy/.venv/lib/python3.7/site-packages/quart/", line 1821, in full_dispatch_request
    result = await self.dispatch_request(request_context)
  File "/usr/lib/rhasspy/.venv/lib/python3.7/site-packages/quart/", line 1869, in dispatch_request
    return await handler(**request_.view_args)
  File "/usr/lib/rhasspy/rhasspy-server-hermes/rhasspyserver_hermes/", line 687, in api_download_profile
  File "/usr/lib/rhasspy/rhasspy-profile/rhasspyprofile/", line 539, in download_files
    f"File size mismatch (got {final_size} byte(s), expected {bytes_expected})", ('', 'File size mismatch (got 20218377 byte(s), expected 20219379)')
[ERROR:2021-04-12 21:52:53,372] Download failure (, got 20218377 byte(s), expected 20219379)
[DEBUG:2021-04-12 21:52:50,395] Downloading to /tmp/tmptdz2subs/tmp8ecno_0i
[DEBUG:2021-04-12 21:52:50,395] (and text_to_speech.system text_to_speech.larynx.vocoder) (larynx vctk_small) ['larynx', 'universal_large'] = False
[DEBUG:2021-04-12 21:52:50,395] (and text_to_speech.system text_to_speech.larynx.vocoder) (larynx vctk_medium) ['larynx', 'universal_large'] = False
[DEBUG:2021-04-12 21:52:50,394] (and text_to_speech.system text_to_speech.larynx.vocoder) (larynx universal_large) ['larynx', 'universal_large'] = True
[DEBUG:2021-04-12 21:52:50,394] (and text_to_speech.system text_to_speech.larynx.default_voice) (larynx thorsten) ['larynx', 'default'] = False
[DEBUG:2021-04-12 21:52:50,394] text_to_speech.system larynx larynx = True
[DEBUG:2021-04-12 21:52:50,394] speech_to_text.deepspeech.mix_weight >0 0 = False
[DEBUG:2021-04-12 21:52:50,394] speech_to_text.kaldi.mix_weight >0 0 = False
[DEBUG:2021-04-12 21:52:50,394] speech_to_text.pocketsphinx.mix_weight >0 0 = False
[DEBUG:2021-04-12 21:52:50,394] speech_to_text.deepspeech.open_transcription True False = False
[DEBUG:2021-04-12 21:52:50,394] speech_to_text.kaldi.open_transcription True False = False
[DEBUG:2021-04-12 21:52:50,393] speech_to_text.pocketsphinx.open_transcription True False = False
[DEBUG:2021-04-12 21:52:50,393] speech_to_text.system deepspeech pocketsphinx = False
[DEBUG:2021-04-12 21:52:50,393] speech_to_text.system kaldi pocketsphinx = False
[DEBUG:2021-04-12 21:52:50,393] Skipping acoustic_model/variances (/profiles/de/acoustic_model/variances)
[DEBUG:2021-04-12 21:52:50,393] Skipping acoustic_model/transition_matrices (/profiles/de/acoustic_model/transition_matrices)
[DEBUG:2021-04-12 21:52:50,393] Skipping acoustic_model/noisedict (/profiles/de/acoustic_model/noisedict)
[DEBUG:2021-04-12 21:52:50,393] Skipping acoustic_model/mixture_weights (/profiles/de/acoustic_model/mixture_weights)
[DEBUG:2021-04-12 21:52:50,393] Skipping acoustic_model/means (/profiles/de/acoustic_model/means)
[DEBUG:2021-04-12 21:52:50,393] Skipping acoustic_model/mdef (/profiles/de/acoustic_model/mdef)
[DEBUG:2021-04-12 21:52:50,392] Skipping acoustic_model/feature_transform (/profiles/de/acoustic_model/feature_transform)
[DEBUG:2021-04-12 21:52:50,392] Skipping acoustic_model/feat.params (/profiles/de/acoustic_model/feat.params)
[DEBUG:2021-04-12 21:52:50,392] Skipping g2p.fst (/profiles/de/g2p.fst)
[DEBUG:2021-04-12 21:52:50,392] Skipping base_dictionary.txt (/profiles/de/base_dictionary.txt)
[DEBUG:2021-04-12 21:52:50,392] speech_to_text.system pocketsphinx pocketsphinx = True

Thanks :smiley:

Would it be possible to just download the file from GitHub and put it there manually to unzip it?

Just copying worked…

For anyone searching in the future, yes you can usually just download and unzip things. Sometimes things need to be put in specific directories. This is controlled entirely within the download section of profile.json.

If you find a file that is split into multiple parts, you can just cat them together and then unzip.