chainerのtrainer.run()がうまく実行できない

前提

http://lovedata.main.jp/2019/06/15/startdeeplearning/
を参考にニューラルネットによる予測モデルを作成しています

実現したいこと

trainerをrunさせたい

発生している問題・エラーメッセージ

Exception in main training loop: 'NN' object has no attribute 'forward'
Traceback (most recent call last):
  File "/usr/local/lib/python3.7/dist-packages/chainer/training/trainer.py", line 343, in run
    update()
  File "/usr/local/lib/python3.7/dist-packages/chainer/training/updaters/standard_updater.py", line 240, in update
    self.update_core()
  File "/usr/local/lib/python3.7/dist-packages/chainer/training/updaters/standard_updater.py", line 253, in update_core
    optimizer.update(loss_func, *in_arrays)
  File "/usr/local/lib/python3.7/dist-packages/chainer/optimizer.py", line 874, in update
    loss = lossfun(*args, **kwds)
  File "/usr/local/lib/python3.7/dist-packages/chainer/link.py", line 286, in __call__
    forward = self.forward  # type: ignore
Will finalize trainer extensions and updater before reraising the exception.

該当のソースコード

python
1#ニューラルネットワークのクラスを作成
2class NN(chainer.Chain):
3
4  #モデルの構造
5  def __init__(self,n_mid_units,n_out=1):
6    super().__init__()
7
8    #パラメータを持つ層の登録
9    with self.init_scope():
10      self.fc1=L.Linear(None,n_mid_units)
11      self.fc2=L.Linear(None,n_out)
12      self.bn=L.BatchNormalization(x_num)
13
14
15      #損失関数の計算
16    def __call__(self,x,t):
17      y=self.predictor(x)
18      loss=F.mean_squared_error(t,y)
19      chainer.report({'loss':loss},self)
20      return loss
21      
22      #順伝搬
23    def predictor(self,x):
24      h=self.bn(x)
25      h=self.fc1(h)
26      h=F.relu(h)
27      h=self.fc2(h)
28      return h
29
30#中間層のノード数、バッチサイズを変化させながら誤差の変化を確認
31for i in [1,3,5,10,20]:
32  print('n_mid_units={}'.format(i))
33
34  #中間層の数を変える
35  model =NN(n_mid_units=i)
36
37  optimizer=chainer.optimizers.Adam()
38  optimizer.setup(model)
39
40  batchsize=10
41  train_iter=chainer.iterators.SerialIterator(train,batchsize)
42  test_iter=chainer.iterators.SerialIterator(train,batchsize,repeat=False,shuffle=False)
43
44  from chainer import training
45  updater=training.StandardUpdater(train_iter,optimizer,device=-1)
46
47  epoch=20
48  trainer=training.Trainer(updater,(20,'epoch'),out='result/electricpower_n_mid_units{}'.format(i))
49  from chainer.training import extensions
50
51  trainer.extend(extensions.Evaluator(test_iter,model,device=-1))
52  trainer.extend(extensions.LogReport(trigger=(1,'epoch')))
53  trainer.extend(extensions.PrintReport(['epoch','main/loss','validation/main/loss','elapsed_time']),trigger=(1,'epoch'))
54