Experiment 4
Binary classification with validation set implementation.
Parameters
basic setting parameters
clip_length: 5.0 # [sec]
preprocessing parameters
sample_rate: 44100
hop_length: 512
n_fft: 1024
n_mels: 64
Training parameters
number of audio samples: 220500
learning rate: 0.001
batch size: 30
number of epochs: 20
number of samples: 72
balanced dataset: False
random clip cut: False
scheduler: True
Classes
Labels
grime
jazz
Class distribution
Dataset | Class | Class ID | Samples |
---|---|---|---|
Train | grime jazz | [0, 1] | 146 |
Validation | grime jazz | [0, 1] | 22 |
Test | grime jazz | [0, 1] | 56 |
Metrics
ID | Epochs | Batch | Classes | LR | samples | Loss | Accuracy |
---|---|---|---|---|---|---|---|
Feb18_21-07-48 | 50 | 73 | 2 | 0.001 | 146 | 0.14 | 0.95 |
Cross entropy
Validation loss is higher than training set. This means the model has gone overfit.
Loss
![](/_next/static/media/Feb18_21-07-48-train-valid-loss.f06e540a.png)
Accuracy
![](/_next/static/media/Feb18_21-07-48-train-valis-acc.6e5a3976.png)
Class Accuracy
![](/_next/static/media/Feb18_21-07-48-class-acc.28e0b8e0.png)
Pr Curves
![](/_next/static/media/Feb18_21-07-48-pr-curves.0ee95a39.png)
Confusion Matrix
![](/_next/static/media/Feb18_21-07-48-matrix.7bf744f8.png)