(model, optimizer, H, X, Y, train_mask)
| 65 | |
| 66 | |
| 67 | def train(model, optimizer, H, X, Y, train_mask): |
| 68 | model.train() |
| 69 | Y_hat = model(H, X) |
| 70 | loss = F.cross_entropy(Y_hat[train_mask], Y[train_mask]) |
| 71 | optimizer.zero_grad() |
| 72 | loss.backward() |
| 73 | optimizer.step() |
| 74 | return loss.item() |
| 75 | |
| 76 | |
| 77 | def evaluate(model, H, X, Y, val_mask, test_mask, num_classes): |