trainingLoss = 0
tp, tn, fp, fn = 0, 0, 0, 0
for _ in range(args.updates_per_epoch):
x, label = next(training_batch)
loss = classifier.update(x, label)
trainingLoss += loss
tpb, tnb, fpb, fnb = classifier.evaluate(x, label)
tp += tpb
tn += tnb
fp += fpb
fn += fnb
trainAcc, trainPrec, _, _, trainFOR = metric.metric(tp, tn, fp, fn)
x_test, label_test = ds.batchTest()
tp, tn, fp, fn = classifier.evaluate(x_test, label_test)
testAcc, testPrec, _, _, testFOR = metric.metric(tp, tn, fp, fn)
trainingLoss /= args.updates_per_epoch
s = "Loss: {: .4f}".format(trainingLoss)
print(epoch, s)
print(
"Training Accuracy = {:.4f}, Precision = {:.4f}, FOR = {:.4f}".format(
trainAcc, trainPrec, trainFOR))
print("Testing Accuracy = {:.4f}, Precision = {:.4f}, FOR = {:.4f}".format(
testAcc, testPrec, testFOR))
if args.save_weights:
classifier.save_weights(args.save_weights)
classifier.done()
