This code has been redesigned to fit 1-D data based on the reference right. Reference
As you can see the results on site Refence2, you can train a high-performance classifier with little labeled data.
10% labeled data | 100% labeled data |
---|---|
0.9255 | 0.945 |
The following result shows the result of classifier learning for 1-D data provided with this code.
As shown in the code below, the classifier(=discriminator) model saved in sgan_flt_diagnosis.py can be loaded and evaluated.
The following result shows the result of classifier test for 1-D data provided with this code.
# load the model
model = load_model('model/c_model_12960.h5')