search

tarekeldeeb / DeepSpeech-Quran

A TensorFlow implementation of Baidu's DeepSpeech architecture
Mozilla Public License 2.0
89 stars 18 forks source link

Tarteel data's gender in deepspeech model #18

Closed smalissa closed 3 years ago

smalissa commented 3 years ago

@tarekeldeeb @aibrahim- @omerasif-itu assalam aliakom pleaze any one help me in the clarification because i have some confusion at some points, if this possible and don’t mind. I want to ask about the following points on Tarteel data:

  1. In Imam model you traind and tested the model on professional resiter dataset , this means that the train data and dev and test have professional records, is this right ?

  2. in Tarteel data ,after you filter them, the dataset that’s used in ( 18420 records), does them include males and females records , or only males records ?.

  3. in Imam + filtered users dataset that’s used: what’s the data used in training set( Tusers only or Imam data only , or mix of Tusers and Imam data ), and in dev set , and also in testing set ? in this step Is this Tusers dataset contains female records or only male records ?.

Thank you in advance.

tarekeldeeb commented 3 years ago

1- your understanding is correct 2- female voices are included 3- dataset is a mix (imam + filtered tarteel users)

Is the methodology clear enough for you?

smalissa commented 3 years ago

@tarekeldeeb @aibrahim- @omerasif-itu thank you for reply and your clarification, but i still have some questions.

  1. How many female records included in the (18,420 ) records were used? I tried to know this but i couldn't.
  3. In “ Imam + Tarteel users “ model, the data of records that’s used is like this:
    • Train CSV file : mix of imam and Tuser records
    • Dev CSV file : mix of imam and Tuser records
      • Test CSV file : mix of imam and Tuser records.
        Is this true ?
  1. Also i need the final CSV files “train, dev, test “ that you used in “Imam + Tuser” model. Can you upload it here or send it to me by email please?

with alot of thanks.

—————————————- Suhad Al-Issa. Computer Engineering. Jordan,Irbid.

From: Tarek Eldeeb @.> Sent: Monday, June 28, 2021 10:29:53 PM To: tarekeldeeb/DeepSpeech-Quran @.> Cc: Suhad Mohammad AL-issa @.>; Author @.> Subject: Re: [tarekeldeeb/DeepSpeech-Quran] Tarteel data's gender in deepspeech model (#18)

Closed #18https://github.com/tarekeldeeb/DeepSpeech-Quran/issues/18.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/tarekeldeeb/DeepSpeech-Quran/issues/18#event-4949405744, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ALUHX2R3FRPDM7XCADYTQRDTVDETDANCNFSM47OPLS5Q.