- Define frozen layers by calling setFeatureExtractor():
MultiLayerNetwork newModel = new TransferLearning.Builder(oldModel)
.setFeatureExtractor(featurizeExtractionLayer)
.build();
- Call fit() to start the training instance:
newModel.fit(numOfEpochs);
MultiLayerNetwork newModel = new TransferLearning.Builder(oldModel)
.setFeatureExtractor(featurizeExtractionLayer)
.build();
newModel.fit(numOfEpochs);