Menu

Error training with TEDLIUM release 1 data

Help
Vamsi
2018-04-19
2018-04-19
  • Vamsi

    Vamsi - 2018-04-19

    Hi,

    I am trying to train a model using TEDLIUM release1 data set and use it with pocketsphinx. However, when i try to run 'sphinxtrain run' command, it fails with the following error

    INFO: sphinx_fe.c(967): Processing all remaining utterances at position 0
    INFO: sphinx_fe.c(787): Converting /Users/vkrish2/Downloads/CMU_Sphinx/TEDLIUM_release1/wav/train_sph/AaronHuey_2010X.sph to /Users/vkrish2/Downloads/CMU_Sphinx/TEDLIUM_release1/feat/train_sph/AaronHuey_2010X.mfc
    ERROR: "sphinx_fe.c", line 204: Input endian big does not match configured value in file '/Users/vkrish2/Downloads/CMU_Sphinx/TEDLIUM_release1/wav/train_sph/AaronHuey_2010X.sph'
    INFO: sphinx_fe.c(787): Converting /Users/vkrish2/Downloads/CMU_Sphinx/TEDLIUM_release1/wav/train_sph/AdamGrosser_2007.sph to /Users/vkrish2/Downloads/CMU_Sphinx/TEDLIUM_release1/feat/train_sph/AdamGrosser_2007.mfc
    ERROR: "sphinx_fe.c", line 204: Input endian big does not match configured value in file '/Users/vkrish2/Downloads/CMU_Sphinx/TEDLIUM_release1/wav/train_sph/AdamGrosser_2007.sph'
    .............

    Can someone please help me out?
    Should I convert sph files from big endian to little endian? If yes, how to do such conversion?
    Thanks in advance.

    Thanks,
    Vamsi

     
    • Nickolay V. Shmyrev

      You need to convert the input to WAV format and split it on utterances.

       

Log in to post a comment.

Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.