'Define by Run'型の深層学習フレームワークが自然言語処理に向いている理由 ref: https://qiita.com/GushiSnow/items/aa660c7228b7024076a8

file0.txt

いい夢見ろよ😴的な? 笑 さっ風呂入ろ ♨ ️
南港に沈めたら解決 
チロルチョコ   
：

file1.txt

いい夢見ろよ😴的な? 笑 さっ風呂入ろ♨ ️
南港に沈めたら解決！<unk><unk><unk> 
チロルチョコ！！<unk><unk><unk><unk>
：

file2.py

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
        self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
        self.conv2_drop = nn.Dropout2d()
        self.fc1 = nn.Linear(320, 50)
        self.fc2 = nn.Linear(50, 10)
        self.conv1_2 = nn.Conv2d(10, 20, kernel_size=5)
        self.conv2_2 = nn.Conv2d(20, 40, kernel_size=5)
        self.fc1_2 = nn.Linear(360, 50)
        self.over_size = 28

    def forward(self, x):
        _, _, h, w = x.size()
        if h > self.over_size:
            x = F.relu(F.max_pool2d(self.conv1(x), 2))
            x = F.relu(F.max_pool2d(self.conv1_2(x), 2))
            x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2_2(x)), 2))
            x = x.view(-1, 360)
            x = F.relu(self.fc1_2(x))
        else:
            x = F.relu(F.max_pool2d(self.conv1(x), 2))
            x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
            x = x.view(-1, 320)
            x = F.relu(self.fc1(x))
        x = F.dropout(x, training=self.training)
        x = self.fc2(x)
        return F.log_softmax(x)