Means of test accurancy and test loss

Question

Friends! I learn "Face Anti-Spoofing using Deep-Pixel-wise-Binary-Supervision". But I have some question because I learn CNN not long time and I can not understand it now. What means test accurancy and test loss below?

def test_accuracy(model, test_dl):
    acc = 0
    total = len(test_dl.dataset)
    for img, mask, label in test_dl:
        net_mask, net_label = model(img)
        preds, _ = predict(net_mask, net_label)
        ac = (preds == label).type(torch.FloatTensor)
        acc += torch.sum(ac).item()
    return (acc / total) * 100


def test_loss(model, test_dl, loss_fn):
    loss = 0
    total = len(test_dl)
    for img, mask, label in test_dl:
        net_mask, net_label = model(img)
        losses = loss_fn(net_mask, net_label, mask, label)
        loss += torch.mean(losses).item()
        #print("test_loss", loss / total)
    return loss / total

I have some results for it

Loss : 0.07998281717300415
Loss : 0.08673418313264847
Loss : 0.10741370171308517
Loss : 0.0790242850780487
Loss : 0.07877358794212341
Loss : 0.09408798813819885
Loss : 0.08567623794078827
Loss : 0.09206777065992355
Loss : 0.08444041013717651
Test Accuracy : 100.0  Test Loss : 0.07839797969375338

Training Complete

Can it be true? Thank you very much!