Menu

Sphinx Train Error at phase 3:

Help
Enigma
2019-06-20
2020-07-26
  • Enigma

    Enigma - 2019-06-20

    Hi I need help
    I using command below

     sphinxtrain

    result below
    Sphinxtrain path: /usr/local/lib/sphinxtrain
    Sphinxtrain binaries path: /usr/local/libexec/sphinxtrain
    Running the training
    MODULE: 000 Computing feature from audio files
    Extracting features from segments starting at (part 1 of 1)
    ERROR: This step had 1 ERROR messages and 0 WARNING messages. Please check the log file for details.
    Extracting features from segments starting at (part 1 of 1)
    ERROR: This step had 1 ERROR messages and 0 WARNING messages. Please check the log file for details.
    Feature extraction is done
    MODULE: 00 verify training files
    Phase 1: Checking to see if the dict and filler dict agrees with the phonelist file.
    Found 20 words using 33 phones
    Phase 2: Checking to make sure there are not duplicate entries in the dictionary
    Phase 3: Check general format for the fileids file; utterance length (must be positive); files exist
    WARNING: Error in '/opt/test/etc/model_test_train.fileids', the feature file '/opt/test/feat/train_si.bd4/TH001_1.mfc' does not exist, or is empty
    Phase 4: Checking number of lines in the transcript file should match lines in fileids file
    WARNING: Transcript file missing a newline at end of file
    Phase 5: Determine amount of training data, see if n_tied_states seems reasonable.
    Estimated Total Hours Training: -2.13675213675214e-07
    This is a small amount of data, no comment at this time
    Phase 6: Checking that all the words in the transcript are in the dictionary
    Words in dictionary: 17
    Words in filler dictionary: 3
    Phase 7: Checking that all the phones in the transcript are in the phonelist, and all phones in the phonelist appear at least once

    What I'm wrong ?

     

    Last edit: Enigma 2019-06-21
  • Enigma

    Enigma - 2019-06-20

    Should I create .fileids and .transcript based on 1st audio file in the main directory ?
    - TH001_1.wav now 1st in audio main directory

     

    Last edit: Enigma 2019-06-20
  • Jake

    Jake - 2019-06-20

    The error message in the log file says

    ERROR: "sphinx_fe.c", line 119: Failed to open /data/SpeechData/train_si.bd4/TH001_1.wav: No such file or directory
    

    You need to firstly follow the error messgae and check if the file is available under the directory.

     
    • Enigma

      Enigma - 2019-06-21

      I have file in directory
      http://imgur.com/xEx097A

       

      Last edit: Enigma 2019-06-21
  • Enigma

    Enigma - 2019-06-21

    Now I know file TH001_1.wav in /data/SpeechData/train_si.bd4/ is just chmod: lrwxrwxrwx
    How do I solve the problem? Or do I have to use only actual files?

     
  • ijazulhassan

    ijazulhassan - 2020-07-26

    I've issue when I start training my model in module:20 "training context independent module" the phase 3 does not exist and it goes to module:60 and goes to completion. After it when I looked into my result and model parameter its empty!!!! @nshmyrev dont know what to do! kindly plz help. @nshmyrev

    MODULE: 20 Training Context Independent models (2020-07-25 20:12)
    Phase 1: Cleaning up directories:

    accumulator... logs... qmanager... models...completed
    Phase 2: Flat initialize

    mk_mdef_gen Log File

    Skipped (set $CFG_CD_TRAIN = 'yes' to enable)

    Skipped (set $CFG_CD_TRAIN = 'yes' to enable)

    Skipped (set $CFG_CD_TRAIN = 'yes' to enable)

    Skipped (set $CFG_CD_TRAIN = 'yes' to enable)

    MODULE: 60 Lattice Generation (2020-07-25 20:12)
    Skipped: $ST::CFG_MMIE set to 'no' in sphinx_train.cfg

    MODULE: 61 Lattice Pruning (2020-07-25 20:12)
    Skipped: $ST::CFG_MMIE set to 'no' in sphinx_train.cfg

    MODULE: 62 Lattice Format Conversion (2020-07-25 20:12)
    Skipped: $ST::CFG_MMIE set to 'no' in sphinx_train.cfg

     

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.