Not logged in.  Login/Logout/Register | List snippets | | Create snippet | Upload image | Upload data

67
LINES

< > BotCompany Repo | #1019262 // TrainVAD

JavaX fragment (include)

sclass TrainVAD {
  replace Recognizer with F1<BWImage, Bool>.
    
  new Best<Recognizer> best;
  LPair<BWImage, Bool> trainingList;
  float stepSizeRatio = 0.25f;
  int clipLength;
    
  Recognizer go() {
    final BWImage img1 = frequencyImage(voiceMegaMix());
    final BWImage img2 = frequencyImage(nonVoiceMegaMix());
    clipLength = iround(2.0*spectro_pixelsPerSecond());
    int stepSize = max(1, iround(clipLength*stepSizeRatio));
    print(+clipLength);
    print(+stepSize);
    L<BWImage> images1 = map(func(IntRange r) -> BWImage { bwHorizontalClip(img1, r) }, stepIntRange(clipLength, intRange(0, img1.getWidth()), stepSize));
    L<BWImage> images2 = map(func(IntRange r) -> BWImage { bwHorizontalClip(img2, r) }, stepIntRange(clipLength, intRange(0, img2.getWidth()), stepSize));
    print(allImageSizes(concatLists(images1, images2)));
    print(l(images1) + " + " + l(images2) + " images");
    
    trainingList = trueFalseBPairs(images1, images2);
    
    // dummy recognizers first
    
    test(func(BWImage img) -> bool { false });
    test(func(BWImage img) -> bool { true });
    
    new AIStrategy_RandomWithVariation<Recognizer> strategy;
    strategy.best = best;
    strategy.submit = f1_scorePreciseRecognizer(trainingList);
    strategy.random = func -> Recognizer {
      nu(VAD_1,
        streakLength := random(1, clipLength/2),
        threshold := randomFloat(),
        band := audio_randomBand())
    };
    strategy.vary = func(Recognizer r) -> Recognizer {
      if (r cast VAD_1)
        ret nu(VAD_1,
          streakLength := varyInt(r.streakLength),
          threshold := varyFloat(r.threshold),
          band := audio_varyBand(r.band));
      null;
    };
    
    runStrategyWhileImprovement(best, strategy);
    
    /*for (int streakLength = 1; streakLength <= clipLength/2; streakLength++)
      test(nu(VAD_1, +streakLength));*/
    
    ret best!;
  }
  
  void test(Recognizer r) {
    scorePreciseRecognizer(best, r, trainingList);
  }
  
  void showErrors(Recognizer r) {
    if (r == null) ret;
    new L<BWImage> errorImages;
    for (Pair<BWImage, Bool> p : trainingList)
      if (r.get(p.a) != p.b)
        errorImages.add(bwAutoContrast(p.a));
    if (nempty(errorImages))
      showImage("Errors", mergeBWImagesVertically(errorImages));
  }
}

download  show line numbers  debug dex  old transpilations   

Travelled to 12 computer(s): bhatertpkbcr, cbybwowwnfue, cfunsshuasjs, gwrvuhgaqvyk, irmadwmeruwu, ishqpsrjomds, lpdgvwnxivlt, mqqgnosmbjvj, pyentgdyhuwx, pzhvpgtvlbxg, tvejysmllsmz, vouqrxazstgt

No comments. add comment

Snippet ID: #1019262
Snippet name: TrainVAD
Eternal ID of this version: #1019262/2
Text MD5: 6fce017840486024fbec2d4a8af56229
Author: stefan
Category: javax / a.i.
Type: JavaX fragment (include)
Public (visible to everyone): Yes
Archived (hidden from active list): No
Created/modified: 2018-10-28 17:44:06
Source code size: 2375 bytes / 67 lines
Pitched / IR pitched: No / No
Views / Downloads: 288 / 820
Version history: 1 change(s)
Referenced in: [show references]