1. # coding=utf-8
    2. import torch
    3. import torch.nn as nn
    4. import torch.nn.functional as F
    5. from torch.autograd import Variable
    8. """Pytorch中神经网络模块化接口nn的了解"""
    9. """
    10. torch.nn是专门为神经网络设计的模块化接口。nn构建于autograd之上，可以用来定义和运行神经网络。
    11. nn.Module是nn中十分重要的类,包含网络各层的定义及forward方法。
    12. 定义自已的网络：
    13.     需要继承nn.Module类，并实现forward方法。
    14.     一般把网络中具有可学习参数的层放在构造函数__init__()中，
    15.     不具有可学习参数的层(如ReLU)可放在构造函数中，也可不放在构造函数中(而在forward中使用nn.functional来代替)
    17.     只要在nn.Module的子类中定义了forward函数，backward函数就会被自动实现(利用Autograd)。
    18.     在forward函数中可以使用任何Variable支持的函数，毕竟在整个pytorch构建的图中，是Variable在流动。还可以使用
    19.     if,for,print,log等python语法.
    21.     注：Pytorch基于nn.Module构建的模型中，只支持mini-batch的Variable输入方式，
    22.     比如，只有一张输入图片，也需要变成 N x C x H x W 的形式：
    24.     input_image = torch.FloatTensor(1, 28, 28)
    25.     input_image = Variable(input_image)
    26.     input_image = input_image.unsqueeze(0)   # 1 x 1 x 28 x 28
    28. """
    31. class LeNet(nn.Module):
    32.     def __init__(self):
    33.         # nn.Module的子类函数必须在构造函数中执行父类的构造函数
    34.         super(LeNet, self).__init__()   # 等价与nn.Module.__init__()
    36.         # nn.Conv2d返回的是一个Conv2d class的一个对象，该类中包含forward函数的实现
    37.         # 当调用self.conv1(input)的时候，就会调用该类的forward函数
    38.         self.conv1 = nn.Conv2d(1, 6, (5, 5))   # output (N, C_{out}, H_{out}, W_{out})`
    39.         self.conv2 = nn.Conv2d(6, 16, (5, 5))
    40.         self.fc1 = nn.Linear(256, 120)
    41.         self.fc2 = nn.Linear(120, 84)
    42.         self.fc3 = nn.Linear(84, 10)
    44.     def forward(self, x):
    45.         x = F.max_pool2d(F.relu(self.conv1(x)), (2, 2))  # F.max_pool2d的返回值是一个Variable
    46.         x = F.max_pool2d(F.relu(self.conv2(x)), (2, 2))
    47.         x = x.view(x.size()[0], -1)
    48.         x = F.relu(self.fc1(x))
    49.         x = F.relu(self.fc2(x))
    50.         x = F.relu(self.fc3(x))
    52.         # 返回值也是一个Variable对象
    53.         return x
    56. def output_name_and_params(net):
    57.     for name, parameters in net.named_parameters():
    58.         print('name: {}, param: {}'.format(name, parameters))
    61. if __name__ == '__main__':
    62.     net = LeNet()
    63.     print('net: {}'.format(net))
    64.     params = net.parameters()   # generator object
    65.     print('params: {}'.format(params))
    66.     output_name_and_params(net)
    68.     input_image = torch.FloatTensor(10, 1, 28, 28)
    70.     # 和tensorflow不一样，pytorch中模型的输入是一个Variable，而且是Variable在图中流动，不是Tensor。
    71.     # 这可以从forward中每一步的执行结果可以看出
    72.     input_image = Variable(input_image)
    74.     output = net(input_image)
    75.     print('output: {}'.format(output))
    76.     print('output.size: {}'.format(output.size()))
    1. net: LeNet(
    2.   (conv1): Conv2d(1, 6, kernel_size=(5, 5), stride=(1, 1))
    3.   (conv2): Conv2d(6, 16, kernel_size=(5, 5), stride=(1, 1))
    4.   (fc1): Linear(in_features=256, out_features=120, bias=True)
    5.   (fc2): Linear(in_features=120, out_features=84, bias=True)
    6.   (fc3): Linear(in_features=84, out_features=10, bias=True)
    7. )
    8. params: <generator object Module.parameters at 0x000001E4BF8F3F20>
    9. name: conv1.weight, param: Parameter containing:
    10. tensor([[[[ 0.0217, -0.1145, -0.1108,  0.0997,  0.1308],
    11.           [-0.0503, -0.0763, -0.1052,  0.0279, -0.1394],
    12.           [-0.1171, -0.0289,  0.1748,  0.1241, -0.0835],
    13.           [ 0.1568,  0.1942, -0.0433,  0.0877,  0.0684],
    14.           [ 0.0543, -0.0093,  0.0070,  0.1326, -0.0348]]],
    17.         [[[ 0.1603, -0.1223,  0.1716,  0.0209, -0.0036],
    18.           [-0.0135,  0.1456, -0.0671,  0.0048, -0.0525],
    19.           [ 0.1955,  0.1027,  0.1439, -0.0273,  0.0086],
    20.           [-0.1991, -0.0868, -0.0525, -0.0175,  0.0813],
    21.           [-0.1898, -0.0448,  0.0030, -0.1157,  0.0471]]],
    24.         [[[ 0.1888,  0.0464,  0.1862,  0.0481,  0.0531],
    25.           [-0.0099, -0.0643, -0.0428,  0.1789,  0.1093],
    26.           [-0.0118,  0.1226, -0.1009, -0.1248, -0.0431],
    27.           [ 0.1237,  0.0433, -0.1344,  0.1017,  0.1664],
    28.           [-0.1594,  0.1199, -0.1465,  0.0877, -0.0816]]],
    31.         [[[-0.0439,  0.1756,  0.1076, -0.0996,  0.1022],
    32.           [-0.0668,  0.1472, -0.0322,  0.1163,  0.1155],
    33.           [ 0.0647,  0.1367,  0.0983,  0.1626, -0.0936],
    34.           [ 0.0148,  0.1392, -0.0226, -0.1734, -0.1294],
    35.           [ 0.0959, -0.0562, -0.0787, -0.0427,  0.1359]]],
    38.         [[[ 0.1670, -0.0294,  0.1815, -0.0586,  0.0111],
    39.           [ 0.1616, -0.0662, -0.1692,  0.1173,  0.0111],
    40.           [ 0.0443, -0.1672,  0.0130, -0.0613, -0.1848],
    41.           [-0.0109,  0.0310,  0.1997, -0.1893, -0.1984],
    42.           [-0.0956, -0.1237,  0.0430,  0.0899,  0.1472]]],
    45.         [[[-0.0851, -0.1191, -0.0242,  0.1510,  0.0598],
    46.           [-0.1764, -0.0255,  0.0431, -0.1236,  0.1123],
    47.           [ 0.1130,  0.0345, -0.0459, -0.1299,  0.0801],
    48.           [ 0.0076, -0.1711,  0.0177, -0.0249,  0.1114],
    49.           [ 0.1724, -0.0620, -0.1699, -0.1931, -0.1063]]]], requires_grad=True)
    50. name: conv1.bias, param: Parameter containing:
    51. tensor([ 0.0458, -0.0054,  0.0085, -0.1519, -0.1371, -0.1475],
    52.        requires_grad=True)
    53. name: conv2.weight, param: Parameter containing:
    54. tensor([[[[-1.6845e-02,  4.4385e-02,  6.8610e-02, -6.5715e-02,  5.6393e-02],
    55.           [-1.5863e-02, -4.3251e-02,  6.7827e-02, -3.4365e-02,  7.6001e-02],
    56.           [-2.6207e-02,  5.5400e-03,  1.7044e-02,  4.7647e-02, -1.9537e-02],
    57.           [-5.0608e-02, -7.4122e-02,  3.3247e-02, -3.1414e-02,  3.2108e-02],
    58.           [-9.5336e-03,  6.9402e-02,  1.4880e-02, -1.8339e-02,  3.6736e-02]],
    60.          [[-2.9151e-03,  6.2530e-02, -2.8998e-02,  6.8531e-02,  3.3358e-02],
    61.           [-2.8497e-02,  4.8150e-02, -2.8942e-02, -1.8290e-02,  2.5050e-02],
    62.           [ 5.8728e-02, -4.1613e-02, -5.0387e-03,  7.1192e-02,  3.7467e-02],
    63.           [ 4.8256e-03, -4.2166e-02, -3.3327e-03,  2.9252e-02,  6.7413e-02],
    64.           [-6.7955e-03,  2.4632e-03,  4.7596e-02, -6.8605e-02,  6.8381e-02]],
    66.          [[-4.9925e-02,  8.0691e-02, -6.1327e-02, -1.4855e-02, -3.7842e-02],
    67.           [-3.4841e-03,  4.9948e-02, -4.3245e-03,  9.5531e-03, -7.8618e-02],
    68.           [-6.7700e-03, -7.3586e-02,  8.0450e-02, -2.4101e-02, -8.0171e-02],
    69.           [-1.5581e-04,  5.4195e-03, -4.3005e-02, -6.7619e-02, -4.0975e-02],
    70.           [ 9.9061e-03,  4.0791e-02, -1.9531e-02,  4.7994e-02,  7.7692e-02]],
    72.          [[-3.3410e-03,  7.0347e-02,  7.9979e-02,  5.4499e-03, -5.6220e-02],
    73.           [-5.6873e-02, -5.3349e-02,  6.2538e-02, -3.7515e-02,  5.6472e-02],
    74.           [-7.2945e-02,  5.6718e-02,  7.8163e-02,  1.7979e-02, -1.0994e-02],
    75.           [ 7.6118e-02,  1.3879e-02,  4.7107e-02,  5.2683e-03,  7.1075e-02],
    76.           [ 4.2975e-02, -5.9535e-02,  1.6814e-02,  2.0818e-02, -5.1213e-02]],
    78.          [[-3.3585e-02, -3.2970e-02,  9.0332e-03, -4.3675e-02,  2.7510e-02],
    79.           [-1.5695e-02, -1.8126e-02,  7.9087e-02, -4.3030e-02,  1.8683e-02],
    80.           [ 3.6872e-03, -5.7530e-02, -1.8339e-02,  3.2536e-02, -9.0900e-03],
    81.           [-3.6027e-02, -3.8068e-02,  1.9422e-02, -3.8830e-02,  3.1912e-02],
    82.           [ 1.5453e-02, -2.4580e-02,  5.2480e-02,  5.2961e-02,  6.2878e-02]],
    84.          [[ 6.7309e-02,  4.3027e-02, -2.1046e-02, -7.1121e-03,  3.5063e-02],
    85.           [-1.1045e-02,  4.5219e-02,  3.2756e-02, -4.9562e-02,  4.8767e-02],
    86.           [ 7.9238e-02, -1.7940e-02,  3.9728e-03, -4.7864e-02,  9.5365e-04],
    87.           [ 7.1904e-02, -2.3580e-02, -2.3010e-03,  4.0284e-02,  7.7537e-02],
    88.           [ 1.0780e-02, -7.2715e-02, -3.5039e-02, -5.6869e-02,  1.1548e-03]]],
    91.         [[[-3.8357e-02, -2.5684e-03, -6.5258e-02, -4.7247e-02,  9.5765e-03],
    92.           [-3.3141e-03, -3.6122e-02,  4.4870e-02,  4.1402e-03,  7.7530e-02],
    93.           [-7.7144e-02,  5.3030e-02, -6.9147e-02,  6.4948e-02, -1.1310e-02],
    94.           [-5.3517e-02,  7.5156e-03, -7.2917e-02, -6.7146e-02, -2.6252e-02],
    95.           [ 3.4591e-02,  5.4645e-02,  5.5110e-02,  1.0251e-02,  4.1073e-02]],
    97.          [[-5.0256e-02,  2.9731e-02,  6.2827e-02, -5.3822e-02,  3.6347e-02],
    98.           [ 6.5142e-02,  3.0691e-02,  5.7352e-02, -3.0507e-02, -1.9708e-02],
    99.           [ 1.0700e-02,  7.9963e-02,  7.2163e-02,  6.1294e-02,  5.3339e-02],
    100.           [-7.4718e-02,  2.5875e-02, -2.8144e-02, -1.2882e-02, -4.0225e-02],
    101.           [ 2.9169e-02, -2.8722e-02, -7.3959e-02, -7.2955e-02,  2.3356e-02]],
    103.          [[-1.4997e-02,  8.7956e-03, -5.0848e-02, -1.9705e-02, -1.0161e-02],
    104.           [-6.8096e-02, -6.8202e-02,  6.4628e-04,  3.5189e-02, -7.5081e-02],
    105.           [ 6.6410e-02,  6.4312e-02, -5.6173e-02,  1.3940e-02,  1.0326e-02],
    106.           [ 1.5539e-03, -4.1628e-03,  5.7444e-02, -5.3338e-02,  1.5603e-02],
    107.           [-4.7347e-02, -2.8402e-02, -2.8120e-02,  7.8141e-03, -6.0936e-02]],
    109.          [[ 1.6892e-02,  2.0420e-02, -5.9399e-02, -2.5154e-03,  3.5316e-02],
    110.           [ 1.1248e-02,  2.4187e-02,  1.3081e-02, -2.4412e-03, -7.3646e-02],
    111.           [ 3.7059e-02,  2.6782e-02, -1.7350e-02,  7.4960e-02, -2.9324e-04],
    112.           [-6.3106e-02, -5.4833e-02,  6.7034e-02, -7.0897e-03, -4.5852e-02],
    113.           [-2.8318e-02, -5.8422e-02,  9.2029e-03,  5.1854e-02, -3.9611e-02]],
    115.          [[ 1.2356e-02, -6.3531e-02,  6.3478e-03,  1.9014e-02,  8.1100e-02],
    116.           [-8.0333e-02,  4.7536e-02,  5.6600e-02, -1.2659e-02, -5.1262e-02],
    117.           [ 1.7703e-02,  3.9341e-02, -2.0800e-02,  4.4321e-02, -2.4239e-02],
    118.           [-2.7211e-02,  6.3447e-02,  2.1908e-02,  1.6242e-02, -4.1113e-02],
    119.           [-7.6609e-03, -1.9560e-02, -5.1269e-02,  3.4129e-02, -5.9217e-02]],
    121.          [[ 6.7518e-02,  6.7900e-02, -7.1994e-02, -1.0922e-02,  1.2612e-02],
    122.           [ 3.3195e-03, -3.8693e-02, -1.1708e-02, -2.0334e-03,  5.4187e-02],
    123.           [ 7.7460e-02, -7.2047e-02,  4.7752e-02,  4.3988e-02, -7.8638e-02],
    124.           [ 4.2541e-02,  1.2595e-02, -4.4230e-02, -3.6403e-02,  4.2550e-02],
    125.           [-2.0077e-02, -6.9298e-02,  4.5013e-02, -6.1551e-02,  9.1950e-03]]],
    128.         [[[-6.3281e-02, -3.7984e-02, -5.2878e-02,  7.5321e-03, -2.8474e-02],
    129.           [ 1.4811e-02,  6.7079e-02,  4.7043e-03, -3.9742e-02, -6.7556e-02],
    130.           [ 3.3260e-03, -3.2924e-02,  6.4261e-02,  6.4207e-02, -1.4801e-02],
    131.           [ 8.0318e-02,  2.5376e-02,  4.6075e-02,  2.2018e-02, -1.0351e-02],
    132.           [-5.5768e-02, -7.0186e-02,  4.6171e-02,  6.2944e-02, -4.8978e-02]],
    134.          [[ 7.2653e-02,  6.1308e-02,  6.6392e-02, -2.2863e-02,  6.9781e-02],
    135.           [-2.7162e-02,  5.5293e-02,  6.3891e-02,  8.9071e-03, -7.7523e-02],
    136.           [ 1.4571e-02, -5.8365e-02, -4.8133e-02,  1.8413e-02,  9.3085e-03],
    137.           [-6.7037e-03, -6.1105e-02, -7.9948e-02, -3.7362e-02,  3.2775e-02],
    138.           [-5.9174e-02,  3.9279e-02,  1.9981e-02, -4.3007e-02,  1.7934e-02]],
    140.          [[-7.7775e-02,  3.6709e-02, -6.7875e-02,  5.3963e-02, -3.1037e-03],
    141.           [-4.8240e-02,  5.7881e-04,  6.0823e-02,  3.9463e-02, -4.7021e-02],
    142.           [ 4.2742e-02, -5.8992e-02,  5.1837e-03, -4.7945e-02, -5.2390e-02],
    143.           [-6.9882e-02, -5.7536e-02, -5.9245e-03, -4.5631e-02, -1.1710e-02],
    144.           [ 2.5742e-02, -6.9348e-02,  4.3669e-02,  6.8487e-02,  2.3740e-03]],
    146.          [[-8.9563e-03, -2.4083e-02, -7.6628e-02,  5.2375e-02,  3.4818e-02],
    147.           [-7.6895e-02,  8.1014e-02, -4.7991e-02, -6.6499e-02,  7.8639e-02],
    148.           [ 3.4462e-02,  5.2886e-02,  7.0614e-02,  4.5503e-04,  2.0397e-02],
    149.           [ 1.8869e-02, -5.6015e-02,  1.6720e-02, -3.3644e-02,  7.6629e-02],
    150.           [-2.7337e-02,  7.9912e-02,  2.9244e-02,  3.7039e-02, -1.3348e-02]],
    152.          [[-5.2761e-02, -5.9233e-02,  5.2979e-02,  5.5442e-02, -6.1234e-02],
    153.           [-2.7945e-02,  7.7628e-02, -7.9512e-02,  3.8591e-02,  7.2761e-02],
    154.           [ 6.9558e-02,  5.5175e-02,  3.0634e-02, -5.1515e-02, -1.7471e-02],
    155.           [-8.0876e-02, -1.6252e-02, -6.9175e-02,  3.5276e-02, -8.0554e-02],
    156.           [-7.8794e-02, -4.0995e-02, -2.8273e-02,  5.8243e-02,  3.0421e-02]],
    158.          [[ 5.9331e-02, -8.4736e-03,  7.9435e-02, -4.3288e-02, -6.9210e-02],
    159.           [-3.8156e-02,  4.2674e-02, -6.5066e-02, -4.1937e-02,  1.2813e-02],
    160.           [-3.2503e-02, -4.6241e-02,  4.1806e-02, -2.4530e-03, -3.7537e-02],
    161.           [-2.6353e-02, -7.2687e-02,  5.8834e-02,  3.0434e-03,  7.4207e-02],
    162.           [-5.1215e-02,  5.7339e-02,  1.5807e-03, -3.5827e-02, -3.4127e-02]]],
    165.         ...,
    168.         [[[ 3.3522e-02,  7.3872e-02, -2.2369e-02, -4.0655e-02, -2.1499e-02],
    169.           [-1.5932e-02, -2.4375e-03,  6.2727e-02, -7.6579e-02,  2.7134e-02],
    170.           [ 6.8823e-02,  6.0422e-02,  6.3404e-04, -5.1311e-02, -9.9124e-03],
    171.           [-6.9576e-02, -2.1669e-02, -4.4207e-02,  3.7925e-02, -3.9863e-02],
    172.           [-3.5746e-02, -2.9237e-03,  1.1273e-02, -5.7513e-02, -5.6617e-02]],
    174.          [[ 5.1629e-02,  7.1810e-02, -5.0015e-02,  7.1614e-03,  3.8804e-02],
    175.           [-7.1633e-02,  6.5081e-02,  5.4332e-02, -3.4122e-02, -2.4071e-02],
    176.           [ 6.3892e-02, -3.2324e-02,  2.3692e-02, -4.3686e-02, -6.8973e-03],
    177.           [ 6.9027e-02, -4.6617e-02,  6.4676e-02, -6.2968e-02,  3.5113e-03],
    178.           [ 5.0808e-02,  2.0932e-03,  3.9056e-02, -3.1558e-02,  3.9545e-02]],
    180.          [[-5.6267e-02, -1.9879e-02,  5.0811e-02,  8.4577e-03, -8.0513e-02],
    181.           [ 2.1660e-02,  6.5831e-02, -4.6980e-02,  3.3146e-02,  6.0543e-02],
    182.           [ 1.2062e-02, -3.7987e-02, -1.5442e-02, -7.7399e-02,  2.1997e-02],
    183.           [-7.2881e-02, -8.0424e-02,  9.0955e-03,  3.4644e-02, -7.5938e-03],
    184.           [-1.1651e-03,  3.8364e-02, -4.1857e-02, -3.9900e-02,  1.1183e-02]],
    186.          [[-6.8551e-02,  3.9157e-02, -1.9457e-04, -9.5278e-03, -4.0466e-02],
    187.           [ 4.4482e-02, -4.7242e-03,  5.2967e-02, -1.6094e-02, -1.1252e-02],
    188.           [-6.4671e-02, -6.9634e-03, -7.1678e-02, -1.8152e-02,  2.5334e-02],
    189.           [-7.6929e-02, -5.2660e-02,  3.8890e-03, -7.1113e-02,  1.5877e-02],
    190.           [ 1.9785e-02,  3.8590e-03,  6.3716e-02, -2.4190e-04,  4.4131e-02]],
    192.          [[-5.6784e-02, -5.0347e-02,  1.5055e-02,  2.7437e-02, -8.0162e-02],
    193.           [-8.7700e-03,  7.1216e-02,  4.7948e-02,  1.8433e-03,  4.5808e-03],
    194.           [ 3.3126e-02, -1.1772e-02, -6.3807e-02, -5.7321e-02, -1.7399e-02],
    195.           [-6.6606e-03,  3.2120e-02,  6.6656e-02,  2.7396e-02, -3.6998e-02],
    196.           [ 4.3725e-02, -2.8585e-02,  9.1075e-03, -7.6035e-02,  3.4320e-02]],
    198.          [[-5.1017e-02,  1.7424e-02,  1.2301e-02,  4.5907e-02, -2.2241e-02],
    199.           [-7.0850e-02, -4.6051e-02,  2.0948e-03, -5.3345e-02,  3.5968e-02],
    200.           [-3.5148e-02, -3.7574e-02, -3.9010e-02,  7.6463e-02, -8.1366e-02],
    201.           [-1.3512e-03,  6.1845e-02, -6.2924e-02,  3.2587e-02, -5.4167e-02],
    202.           [ 7.7570e-02, -7.9204e-02,  4.1169e-02,  6.8774e-02, -1.2836e-02]]],
    205.         [[[-7.0554e-02,  2.7973e-02,  5.9153e-02, -7.2539e-02,  3.1440e-02],
    206.           [-7.5312e-02,  1.1987e-02,  7.5666e-02,  2.3965e-03,  1.4349e-02],
    207.           [-5.5436e-02,  3.1992e-02, -1.4974e-02,  5.4549e-03, -1.8292e-02],
    208.           [ 1.4059e-02,  8.4250e-03, -2.1723e-02,  7.6211e-02,  5.3961e-02],
    209.           [ 5.6001e-03, -8.0475e-02,  6.9256e-03,  4.9242e-02, -3.5211e-02]],
    211.          [[-2.2841e-02,  4.3061e-02, -5.6875e-02, -7.5937e-02, -6.0857e-02],
    212.           [-5.9746e-02,  3.4754e-03, -2.5920e-02, -7.0835e-02,  2.8814e-03],
    213.           [-6.9817e-02,  1.8242e-02,  5.0231e-02, -3.8049e-02, -6.4146e-02],
    214.           [-5.3743e-03,  1.1317e-02,  1.2650e-02,  6.6841e-02,  1.3782e-02],
    215.           [ 4.0454e-02, -1.7964e-02, -2.1426e-03, -3.0909e-02,  3.9788e-02]],
    217.          [[-7.6786e-02, -2.4914e-02,  6.0819e-02,  4.7445e-02, -6.5816e-02],
    218.           [-5.0722e-02,  1.9875e-02,  6.7958e-02,  5.0013e-02, -1.7848e-02],
    219.           [-4.6331e-02,  7.7802e-02,  6.2882e-02, -7.9509e-02,  2.8542e-02],
    220.           [-4.2428e-02, -4.7086e-02,  4.4445e-02, -7.9303e-02, -4.7935e-02],
    221.           [ 6.3782e-02,  5.0679e-02, -3.7921e-02, -1.9698e-02,  5.0653e-02]],
    223.          [[-7.0076e-03,  4.9157e-02,  7.4267e-02, -7.8928e-02,  4.0009e-02],
    224.           [-1.8811e-02, -2.0486e-03, -1.5589e-02, -3.4986e-02, -3.7412e-02],
    225.           [ 1.6700e-02,  4.1947e-06,  2.6909e-03, -1.3102e-02,  5.3462e-02],
    226.           [-2.9426e-02, -3.0296e-02,  6.4140e-03,  4.4588e-02, -7.4350e-02],
    227.           [-4.5037e-02, -3.1994e-02,  1.2310e-02,  8.1318e-02,  1.9178e-02]],
    229.          [[ 6.1147e-02,  3.0579e-02,  7.2085e-02, -2.2600e-02,  5.6355e-02],
    230.           [-2.1143e-02, -4.7402e-02,  6.6499e-02, -5.2549e-02,  5.5194e-02],
    231.           [-6.9439e-02, -5.3459e-02, -7.2381e-02,  9.7566e-03,  8.4158e-03],
    232.           [-8.0540e-02, -2.0421e-02, -3.1750e-02, -5.5141e-02, -6.0945e-03],
    233.           [ 6.8186e-02, -6.6666e-02,  1.1205e-02, -2.0387e-02,  4.4871e-04]],
    235.          [[-1.4830e-02, -2.2399e-02, -5.5202e-03, -3.9772e-02, -7.9892e-02],
    236.           [ 3.6414e-02, -4.2354e-02,  5.7111e-02,  2.3401e-02,  4.9164e-02],
    237.           [-1.5300e-03,  1.2318e-02,  3.8444e-02, -6.6135e-02,  1.1690e-02],
    238.           [-8.0992e-02, -5.7289e-02, -4.2210e-02,  4.7406e-02,  1.7448e-02],
    239.           [-7.8387e-02, -4.2730e-02, -3.4001e-03,  2.2935e-02,  7.0695e-02]]],
    242.         [[[-5.4475e-02, -5.3104e-02, -3.1606e-02, -6.6535e-02,  6.6955e-02],
    243.           [-3.8603e-02, -1.4596e-02,  4.4006e-02,  7.6619e-02, -6.5071e-02],
    244.           [ 5.0730e-02, -4.5500e-02, -4.5379e-02,  1.9727e-02,  2.3672e-02],
    245.           [-1.5720e-02, -6.8718e-02,  1.5912e-02, -4.5637e-02,  1.2738e-02],
    246.           [ 1.3480e-02, -5.9709e-02, -5.5168e-02, -2.2764e-02,  1.4727e-02]],
    248.          [[-3.9977e-02, -6.6003e-02, -5.8183e-03, -1.8981e-02, -5.9466e-02],
    249.           [-8.1542e-02,  7.2437e-02,  7.2222e-02,  7.1420e-02, -3.1859e-02],
    250.           [-2.0119e-02, -7.9562e-02,  3.9553e-02,  3.4676e-02,  4.8136e-02],
    251.           [-2.9620e-02, -5.7869e-02,  2.8632e-02, -5.4495e-02,  2.9654e-02],
    252.           [ 7.5934e-02,  5.5229e-02,  6.0961e-02, -7.0632e-02, -5.8143e-02]],
    254.          [[-4.4237e-02, -6.4266e-03, -3.8340e-02,  7.9527e-02,  2.7897e-02],
    255.           [ 7.7764e-02,  1.6796e-03,  6.2879e-02,  2.9104e-02, -2.2573e-02],
    256.           [-5.2889e-02,  4.8257e-02, -4.0350e-02,  1.6663e-02, -1.9207e-02],
    257.           [-6.0075e-02,  2.0691e-02,  2.0772e-02,  5.5525e-02,  7.0949e-02],
    258.           [ 2.7106e-02, -7.3379e-02,  6.2820e-02, -7.0091e-02, -1.0926e-02]],
    260.          [[-2.7131e-02, -2.0585e-02, -7.5044e-03, -7.2443e-03, -8.2568e-03],
    261.           [-6.5148e-02,  7.3872e-02,  6.6669e-02,  6.4592e-02, -5.5050e-02],
    262.           [-4.6178e-02, -2.7139e-03, -4.4908e-02,  6.4572e-02, -4.4041e-02],
    263.           [-8.1407e-02, -1.7506e-02,  6.0459e-02, -4.2148e-02,  2.5116e-02],
    264.           [ 5.9035e-02,  4.1849e-02, -5.5773e-02, -5.0496e-02, -9.0559e-03]],
    266.          [[ 6.0382e-02, -8.0480e-02,  5.4642e-03,  1.5028e-02,  2.1725e-02],
    267.           [-5.3438e-02, -5.1294e-02, -1.8862e-02, -1.7900e-04, -5.7885e-02],
    268.           [ 5.5422e-02,  1.7491e-02,  2.5857e-02,  6.9776e-02, -5.3682e-02],
    269.           [ 1.4133e-02,  6.1083e-02,  2.9480e-02,  2.6013e-02,  3.3310e-02],
    270.           [-8.0320e-02,  2.8635e-02, -6.0207e-02,  4.1884e-02, -1.4888e-04]],
    272.          [[-5.0174e-02, -1.4831e-02,  1.3534e-02, -4.3973e-02, -3.6332e-02],
    273.           [ 6.1421e-02, -2.9783e-02,  7.4678e-02, -4.5586e-02, -1.9699e-02],
    274.           [-6.0944e-02, -4.6267e-03,  5.1604e-02, -4.4676e-02,  4.7587e-02],
    275.           [-1.9579e-02,  7.3082e-03,  7.1702e-02,  6.8381e-02,  3.1841e-02],
    276.           [ 7.1288e-02,  5.2930e-02,  7.0572e-02, -1.8718e-02,  6.3015e-02]]]],
    277.        requires_grad=True)
    278. name: conv2.bias, param: Parameter containing:
    279. tensor([-0.0782, -0.0083,  0.0646,  0.0656, -0.0500, -0.0608, -0.0603, -0.0301,
    280.         -0.0017,  0.0806, -0.0535, -0.0409,  0.0495, -0.0180,  0.0180,  0.0131],
    281.        requires_grad=True)
    282. name: fc1.weight, param: Parameter containing:
    283. tensor([[ 0.0557,  0.0568, -0.0609,  ...,  0.0324, -0.0108, -0.0446],
    284.         [ 0.0421, -0.0183,  0.0237,  ...,  0.0613, -0.0095, -0.0477],
    285.         [ 0.0229,  0.0344, -0.0473,  ..., -0.0393,  0.0215,  0.0060],
    286.         ...,
    287.         [ 0.0111,  0.0099, -0.0386,  ...,  0.0008, -0.0354,  0.0485],
    288.         [-0.0620,  0.0587, -0.0312,  ...,  0.0541, -0.0224, -0.0386],
    289.         [-0.0129,  0.0024, -0.0270,  ...,  0.0153, -0.0243,  0.0266]],
    290.        requires_grad=True)
    291. name: fc1.bias, param: Parameter containing:
    292. tensor([-0.0590,  0.0625, -0.0362, -0.0115,  0.0260,  0.0553,  0.0619,  0.0085,
    293.          0.0023, -0.0315, -0.0234, -0.0243,  0.0522, -0.0514, -0.0554, -0.0047,
    294.          0.0296, -0.0058, -0.0614, -0.0430,  0.0461,  0.0293, -0.0249, -0.0037,
    295.          0.0505, -0.0068, -0.0399, -0.0603, -0.0063,  0.0344, -0.0551,  0.0420,
    296.          0.0152,  0.0062, -0.0392, -0.0251,  0.0350,  0.0392,  0.0239,  0.0539,
    297.          0.0602, -0.0601, -0.0253, -0.0094, -0.0086,  0.0301,  0.0579, -0.0293,
    298.         -0.0530, -0.0157, -0.0089,  0.0089,  0.0350,  0.0260, -0.0440, -0.0153,
    299.         -0.0492, -0.0111, -0.0303, -0.0462,  0.0400, -0.0418, -0.0290,  0.0115,
    300.          0.0207,  0.0108, -0.0083,  0.0334,  0.0065,  0.0317,  0.0048,  0.0122,
    301.         -0.0270, -0.0186,  0.0565,  0.0032,  0.0081,  0.0060, -0.0493,  0.0466,
    302.          0.0260,  0.0325,  0.0040, -0.0083,  0.0017,  0.0591, -0.0496, -0.0103,
    303.          0.0230,  0.0466, -0.0498,  0.0320,  0.0531, -0.0411, -0.0324, -0.0005,
    304.          0.0315, -0.0126,  0.0217,  0.0043,  0.0047,  0.0578, -0.0498, -0.0359,
    305.         -0.0142,  0.0498, -0.0506, -0.0491, -0.0336,  0.0014, -0.0323,  0.0126,
    306.         -0.0531,  0.0186, -0.0277, -0.0394, -0.0089, -0.0005, -0.0092,  0.0531],
    307.        requires_grad=True)
    308. name: fc2.weight, param: Parameter containing:
    309. tensor([[ 0.0382, -0.0780,  0.0119,  ...,  0.0212,  0.0506, -0.0287],
    310.         [ 0.0372,  0.0032,  0.0497,  ...,  0.0545,  0.0307, -0.0403],
    311.         [ 0.0225,  0.0167,  0.0551,  ..., -0.0497, -0.0162, -0.0655],
    312.         ...,
    313.         [-0.0686,  0.0433, -0.0684,  ..., -0.0263, -0.0782, -0.0531],
    314.         [-0.0701, -0.0892,  0.0189,  ...,  0.0175,  0.0305,  0.0491],
    315.         [-0.0664,  0.0643,  0.0743,  ...,  0.0454, -0.0245, -0.0781]],
    316.        requires_grad=True)
    317. name: fc2.bias, param: Parameter containing:
    318. tensor([ 0.0119, -0.0211,  0.0746,  0.0071, -0.0032, -0.0312,  0.0082, -0.0478,
    319.         -0.0862, -0.0892, -0.0274, -0.0791, -0.0009,  0.0488,  0.0253,  0.0011,
    320.          0.0037,  0.0067, -0.0903, -0.0320, -0.0522, -0.0800, -0.0606,  0.0886,
    321.          0.0281,  0.0756,  0.0606,  0.0911, -0.0089,  0.0742, -0.0669,  0.0517,
    322.          0.0721, -0.0175,  0.0174, -0.0802,  0.0600,  0.0308, -0.0065,  0.0591,
    323.          0.0571,  0.0242, -0.0801,  0.0246, -0.0314,  0.0472, -0.0414,  0.0452,
    324.         -0.0297,  0.0519, -0.0375, -0.0871, -0.0716,  0.0212,  0.0266, -0.0608,
    325.          0.0515,  0.0055,  0.0823,  0.0249, -0.0304, -0.0548,  0.0034,  0.0372,
    326.          0.0640,  0.0609, -0.0632, -0.0131, -0.0469,  0.0890,  0.0256, -0.0546,
    327.          0.0343,  0.0109, -0.0718, -0.0138, -0.0912,  0.0312,  0.0509, -0.0576,
    328.         -0.0757, -0.0018, -0.0336,  0.0336], requires_grad=True)
    329. name: fc3.weight, param: Parameter containing:
    330. tensor([[ 2.5675e-02,  9.6464e-02, -2.8338e-02, -3.7553e-02,  8.3534e-02,
    331.           6.1218e-02,  7.5608e-02, -7.0195e-02,  1.3130e-02, -7.2840e-02,
    332.          -8.2775e-02, -8.6803e-02, -3.0200e-02,  4.9099e-02,  5.6712e-02,
    333.           2.7599e-02, -1.0273e-02,  7.4255e-02,  4.2497e-02,  5.5611e-02,
    334.           2.7333e-02,  1.0501e-01, -2.5311e-02, -4.7043e-02, -8.8774e-02,
    335.           6.3224e-02,  7.9647e-02,  6.5620e-02, -7.2866e-02,  1.0637e-01,
    336.           3.3025e-02,  3.6889e-02,  9.9651e-02,  5.7380e-02,  2.4613e-02,
    337.           3.5757e-02, -3.8497e-02,  6.0118e-03,  1.0798e-01, -1.6243e-02,
    338.          -9.6955e-02, -4.5885e-02,  6.7519e-02, -4.3009e-02,  3.9104e-02,
    339.           4.6458e-02, -5.7090e-02, -1.0370e-01,  4.5300e-02, -1.0485e-01,
    340.           6.1116e-02,  7.6842e-02,  9.8063e-02,  6.8675e-02, -1.0554e-01,
    341.           2.0175e-02,  7.2442e-02, -9.5526e-02, -9.5997e-02, -1.0504e-01,
    342.          -9.4131e-02,  6.1151e-02,  9.9820e-02,  8.9036e-02, -2.0200e-02,
    343.          -7.2124e-02,  9.3611e-02,  1.0573e-01, -1.4734e-04, -1.0717e-01,
    344.           2.2878e-02, -2.4432e-02, -1.0675e-01, -3.2646e-02, -5.8235e-02,
    345.           5.9726e-02,  5.3219e-02, -3.1837e-02,  3.3325e-02,  5.5540e-02,
    346.           8.1466e-02,  9.1072e-02, -5.8518e-02, -3.1267e-02],
    347.         [ 8.2630e-02, -8.3926e-02, -2.1073e-02,  6.0609e-02,  1.0743e-01,
    348.          -1.0275e-01,  9.0790e-02, -2.8891e-02, -4.3349e-02,  3.1871e-02,
    349.          -5.5935e-03, -1.7263e-02,  4.6537e-02,  6.9056e-02,  1.0890e-01,
    350.           7.9869e-02, -3.3150e-02, -5.6133e-02,  7.4652e-02,  1.5729e-02,
    351.           7.7521e-02,  1.7651e-02,  6.5500e-02,  1.1925e-02, -7.8438e-02,
    352.          -7.5975e-02,  1.0549e-02,  1.0411e-01, -7.8864e-02,  3.3789e-02,
    353.          -6.8366e-02, -9.9897e-02, -5.9828e-02,  4.6674e-02,  8.8915e-02,
    354.          -4.7183e-02,  2.9351e-02, -1.0050e-01, -3.0687e-02, -8.7500e-02,
    355.          -1.4717e-02, -9.0385e-02, -1.8436e-02, -4.1140e-02,  5.8814e-02,
    356.           5.8907e-02,  7.8952e-02,  1.7692e-02, -2.4836e-02,  6.5768e-02,
    357.          -4.1007e-02,  4.1342e-03,  7.6616e-02, -8.2241e-03, -2.6829e-02,
    358.          -4.7358e-02,  1.1062e-02, -7.5820e-02, -7.5925e-02,  6.0198e-02,
    359.          -5.4272e-02, -1.3632e-02,  1.0298e-02,  1.0168e-01,  7.0716e-02,
    360.          -2.1892e-02,  8.7976e-02,  1.8015e-02, -1.9524e-02,  9.7254e-02,
    361.           6.3282e-02,  3.1983e-03, -5.2951e-02,  9.3937e-02,  3.3288e-02,
    362.           4.0928e-02,  2.4922e-02, -4.7823e-02,  8.6050e-02, -8.4335e-02,
    363.           4.7266e-02, -6.1726e-02, -1.0203e-01,  8.8243e-02],
    364.         [ 4.1489e-02, -9.1939e-02,  6.8136e-02,  3.5553e-02, -9.0968e-02,
    365.          -9.3154e-02, -3.1098e-02,  3.1214e-03, -9.0762e-02, -1.8996e-02,
    366.           9.3973e-02,  2.0913e-02,  4.7864e-02,  9.3036e-02, -9.0032e-02,
    367.          -6.5305e-02,  1.0673e-01,  7.4308e-02, -3.8425e-02,  9.5343e-02,
    368.           2.0384e-02, -1.0299e-01, -8.2269e-02, -9.8493e-02, -7.9151e-02,
    369.           9.1867e-02,  2.7711e-03,  4.2344e-02,  7.2716e-02, -8.8083e-02,
    370.           5.7648e-02, -3.1896e-02,  7.5422e-02, -6.8545e-02, -6.8769e-02,
    371.           6.1481e-02, -1.5343e-02,  7.8233e-02,  6.4081e-02,  1.0791e-01,
    372.           2.8595e-02, -9.1479e-02,  8.7110e-03, -4.5246e-02, -2.1293e-02,
    373.           1.0427e-01,  6.3909e-02,  1.0082e-02, -2.4838e-02,  6.1349e-02,
    374.          -2.3162e-02,  8.3248e-02, -6.8129e-02,  4.5732e-02,  4.3183e-02,
    375.           9.0050e-02, -9.8949e-02,  7.9788e-02, -4.1124e-02, -2.9742e-02,
    376.           9.0467e-02,  5.8855e-02,  9.6743e-02, -4.6616e-02,  6.4971e-02,
    377.          -1.0466e-01, -6.3834e-02,  9.7982e-02, -1.0604e-01,  1.9148e-02,
    378.          -1.4930e-02,  3.8111e-02, -5.2557e-02, -7.9686e-02, -6.6230e-02,
    379.          -6.5741e-02, -1.5909e-02, -1.0386e-01,  5.5472e-02,  6.7832e-02,
    380.           6.3233e-02,  7.9981e-02,  9.8060e-02,  8.3838e-02],
    381.         [ 7.0022e-02,  9.0171e-02,  5.9061e-02,  2.1166e-02,  8.8820e-02,
    382.          -7.8665e-02,  8.8208e-02,  8.5584e-02,  7.4943e-02, -7.7087e-02,
    383.           8.5376e-02,  3.3827e-02, -8.3241e-02,  2.1999e-04, -2.8975e-03,
    384.          -6.9084e-02,  1.6488e-02,  5.7707e-02,  1.8457e-02, -4.5824e-02,
    385.           5.4020e-02, -9.6926e-02, -4.2203e-02, -2.1521e-02, -3.7959e-02,
    386.          -1.5385e-02,  7.8936e-02, -1.0864e-01,  6.7912e-02, -9.6278e-02,
    387.           7.1233e-02, -2.8165e-02, -1.0583e-01, -9.6868e-02, -6.4872e-02,
    388.          -5.9130e-02,  1.0399e-01, -1.0651e-04,  7.3894e-02,  4.0169e-03,
    389.           1.0359e-01, -7.2857e-02, -4.7068e-02,  5.0677e-02, -3.7540e-02,
    390.           7.5862e-02, -2.7972e-02,  3.4016e-02, -4.4954e-02, -9.7638e-02,
    391.          -1.0810e-01,  5.2891e-02,  2.9980e-02,  5.0854e-02,  5.4759e-03,
    392.           8.7307e-02,  6.1735e-02,  4.4016e-02, -4.5422e-02, -2.8660e-02,
    393.          -7.9815e-02, -2.5486e-02,  1.1003e-02,  1.6672e-02, -1.0703e-01,
    394.          -9.3843e-03,  7.1490e-02, -2.9863e-02,  5.6345e-02,  3.0910e-02,
    395.           2.6883e-02,  5.8811e-02, -6.1727e-02, -3.7157e-02, -1.6888e-02,
    396.          -8.9994e-02, -9.3659e-02, -4.6373e-02,  1.5054e-02,  2.0068e-02,
    397.          -6.0038e-02,  4.2358e-02,  6.3696e-02, -4.3962e-02],
    398.         [-8.3121e-02,  8.5531e-02, -2.6016e-02, -6.8964e-02,  4.5352e-02,
    399.           1.5961e-02,  5.8868e-02, -2.5308e-02,  7.6926e-02, -6.8305e-02,
    400.           1.8905e-02, -5.8907e-02,  1.0423e-01,  4.9885e-02,  9.5187e-02,
    401.          -7.1909e-02,  3.4269e-02, -3.9465e-02, -2.9066e-02, -8.0808e-02,
    402.           8.8203e-02,  4.3563e-02,  1.5458e-02, -5.3114e-02, -1.0442e-01,
    403.          -4.3608e-02,  2.4528e-02, -2.9076e-02, -1.0458e-01, -1.7147e-03,
    404.           3.0054e-03, -8.3290e-03, -6.9130e-03, -8.5197e-02,  5.3025e-02,
    405.          -9.6852e-03, -5.5586e-02, -2.2490e-02,  5.9178e-02, -1.0007e-01,
    406.           6.3793e-02, -2.8858e-02, -1.6784e-02, -9.3244e-02,  2.9719e-02,
    407.          -6.7288e-03, -7.7317e-02,  4.8673e-02,  5.0237e-02, -7.6395e-02,
    408.          -2.0794e-02, -6.3980e-02, -4.3560e-02, -8.7049e-03,  6.4666e-02,
    409.          -1.3038e-02, -4.5066e-03, -5.6956e-02,  6.7503e-02,  3.5688e-02,
    410.          -9.5214e-02,  4.1263e-02,  1.5851e-02,  6.4534e-02,  2.3171e-02,
    411.          -4.9850e-02, -9.1811e-02,  6.2961e-02, -3.4310e-02, -6.2068e-02,
    412.          -1.6055e-02, -1.3338e-02,  6.0653e-02, -5.7536e-02,  2.6076e-02,
    413.           8.9688e-03,  8.3964e-02, -9.3622e-02, -1.0663e-01, -5.6277e-02,
    414.          -5.8059e-02, -1.3131e-02, -3.8715e-02,  6.2020e-02],
    415.         [-2.6569e-02,  5.4337e-02, -9.6171e-02, -5.5391e-02, -4.9317e-02,
    416.          -2.6928e-03,  9.7091e-02, -4.0224e-02,  1.0825e-01,  6.7988e-02,
    417.          -9.7152e-02, -2.6146e-02,  7.5801e-02, -9.9237e-02, -2.5318e-02,
    418.           1.0640e-01, -2.7331e-02,  5.2327e-02, -3.9254e-02,  9.3247e-03,
    419.          -9.5941e-02,  9.6498e-03,  5.4255e-02, -8.7857e-02,  6.8306e-03,
    420.           2.7035e-03, -4.6665e-02,  1.3167e-02, -5.1213e-02,  4.9794e-02,
    421.          -2.7187e-02, -8.7509e-02, -8.5662e-02, -6.9200e-02,  6.6691e-02,
    422.          -5.0088e-02, -2.9523e-02,  1.5589e-03,  9.0469e-02, -3.9184e-02,
    423.           8.5950e-03,  4.2269e-02, -5.4336e-02,  8.1889e-02,  1.0891e-01,
    424.           5.8277e-03, -6.4641e-02,  7.9374e-03,  7.7086e-03, -1.0774e-01,
    425.           4.5939e-02,  3.7619e-02,  9.8091e-02, -5.1942e-02, -8.9840e-02,
    426.          -6.4495e-02,  4.6240e-03,  4.7031e-02, -3.3025e-02, -4.2144e-02,
    427.           8.3233e-02, -9.0001e-02, -2.5094e-02,  1.0504e-01,  2.5988e-02,
    428.           2.1935e-02, -3.7314e-02, -2.8549e-02, -6.8368e-02, -8.0791e-02,
    429.          -3.3495e-02,  8.7517e-02,  9.6160e-02,  9.9490e-02,  1.2337e-02,
    430.          -9.0141e-02,  9.5914e-02,  8.1420e-02,  1.0717e-01,  1.0881e-01,
    431.           8.7895e-02, -7.3054e-02,  7.2266e-02,  6.9090e-03],
    432.         [-8.0899e-02, -5.2218e-02, -1.0036e-01, -8.2043e-02,  1.0096e-01,
    433.          -4.2331e-02, -7.5092e-02,  3.1133e-02,  9.2855e-03,  8.3895e-02,
    434.           6.7684e-02,  4.4834e-02, -1.0125e-01, -4.6818e-02,  7.5683e-02,
    435.          -2.7669e-02, -8.4572e-02,  3.8523e-03,  6.1291e-02,  9.1819e-02,
    436.          -4.2794e-02, -9.0245e-03,  6.4425e-02, -8.1464e-02, -9.3352e-02,
    437.          -4.7467e-02, -8.6843e-02, -8.6398e-02, -4.3590e-02, -7.1583e-02,
    438.           5.1992e-02, -7.9497e-02,  3.5019e-02,  1.0563e-01, -2.5732e-02,
    439.           9.9013e-02,  4.3581e-02,  4.4501e-02,  1.8952e-03,  3.3189e-02,
    440.           3.7902e-02, -6.4136e-02,  8.1677e-02,  5.0096e-02, -6.2569e-02,
    441.           1.1812e-02, -9.6668e-02,  1.0521e-01, -5.8639e-02, -1.0639e-01,
    442.           3.1409e-02, -1.3698e-02, -3.9796e-02,  5.1740e-02, -5.5802e-02,
    443.           4.8692e-02, -4.1490e-02,  1.1403e-02, -4.6092e-02, -3.6024e-02,
    444.          -1.1991e-02,  7.1447e-02,  4.4755e-02, -8.4097e-02,  5.0799e-02,
    445.           3.6624e-02,  6.7349e-02,  3.4251e-02, -3.8027e-02,  5.5241e-02,
    446.          -6.5773e-03, -6.6297e-02, -5.1774e-02,  1.6591e-02, -4.6628e-02,
    447.           8.4549e-02,  3.5520e-02, -4.2439e-03, -1.1652e-02,  3.7523e-03,
    448.           6.7728e-02, -9.0507e-02,  7.2635e-02, -4.5230e-02],
    449.         [ 2.9167e-02,  5.4838e-03, -4.5849e-02, -1.8952e-02, -5.2089e-02,
    450.           3.8475e-02, -3.9655e-03,  3.8433e-02,  9.0695e-02, -1.9578e-02,
    451.          -7.4018e-02,  4.3374e-04,  2.4131e-02,  8.1184e-02,  2.8429e-02,
    452.          -8.4612e-03,  2.1370e-03,  7.9961e-02,  6.6877e-02, -6.3081e-02,
    453.           1.0406e-01, -1.5700e-02,  4.2943e-02, -2.4016e-02,  1.8705e-02,
    454.           5.1955e-02, -8.5558e-02, -1.0711e-01, -5.2827e-02,  4.3816e-02,
    455.          -2.9981e-02,  9.3320e-02,  6.5789e-02, -9.1292e-03,  9.1734e-02,
    456.          -1.0477e-01, -8.7946e-02, -6.3073e-02, -6.4625e-02,  3.3974e-02,
    457.           9.1596e-02, -1.0346e-01,  3.0706e-03, -5.9908e-02,  7.0668e-02,
    458.          -7.8272e-02,  5.4159e-02, -5.9159e-02,  2.4638e-02,  9.5201e-02,
    459.           9.7090e-02, -2.6794e-03,  1.3644e-02,  8.6832e-02, -1.0815e-01,
    460.           3.1548e-02,  1.0873e-01, -6.0625e-04,  7.9496e-02,  3.9644e-02,
    461.          -9.5759e-02,  4.5270e-02, -5.7831e-02, -2.4183e-02, -4.0251e-02,
    462.          -2.2646e-02,  7.5958e-02,  3.1137e-02, -2.9679e-02, -8.5065e-02,
    463.           6.0740e-02,  2.6547e-02,  3.7758e-02,  1.9960e-02,  7.5465e-02,
    464.           1.3742e-02,  4.1609e-02, -7.5583e-02, -7.4171e-02,  4.1864e-02,
    465.          -8.4194e-02,  4.6219e-02, -4.4277e-02,  7.3752e-02],
    466.         [ 8.6970e-02, -8.2441e-05,  9.3113e-03, -4.5907e-02,  1.0563e-01,
    467.          -8.7527e-02,  1.0711e-02, -5.9712e-02, -3.4388e-02,  5.4009e-02,
    468.           8.0843e-03,  5.3925e-02, -1.0245e-01,  7.0390e-02, -6.3553e-02,
    469.          -5.7141e-02, -8.6154e-02, -6.7943e-03,  6.3098e-02, -8.9417e-02,
    470.          -8.5114e-02, -5.5248e-02, -1.0754e-02, -8.1079e-02, -2.5598e-02,
    471.          -5.7460e-02, -1.9243e-02, -1.2445e-02,  3.6473e-02,  6.4427e-02,
    472.          -5.8915e-02,  9.9241e-02,  9.3676e-02,  7.9765e-02,  1.0666e-01,
    473.          -2.0676e-02, -5.0233e-02,  1.0802e-01,  7.9624e-02, -6.4746e-02,
    474.          -3.5036e-02,  6.2020e-02, -6.5206e-02, -8.2924e-02, -3.9141e-02,
    475.          -4.8113e-03, -3.3400e-03,  7.2341e-02, -5.8601e-02, -1.0542e-01,
    476.           1.0210e-01, -9.9237e-02, -8.4985e-02, -4.3769e-02,  1.8451e-02,
    477.           6.8206e-02,  8.2744e-02, -8.0338e-02,  9.6329e-02,  7.5431e-03,
    478.          -1.2411e-02, -8.4952e-02,  7.8804e-02, -3.4132e-02, -7.9321e-02,
    479.           5.4848e-02,  2.0019e-02, -5.5738e-02,  1.0282e-01, -9.8613e-02,
    480.           7.6550e-03, -3.2091e-02, -9.7843e-02, -2.1994e-02,  6.7543e-02,
    481.          -7.1321e-03, -8.2259e-03, -1.9674e-02,  2.6065e-02,  6.9857e-02,
    482.          -9.4976e-02,  5.6883e-02, -4.2162e-02, -5.2965e-02],
    483.         [-9.6068e-02,  3.1667e-02, -8.7749e-02, -6.3179e-02,  7.4492e-02,
    484.           8.9063e-02, -3.3481e-02, -2.4020e-02, -1.3834e-02, -2.4226e-02,
    485.           8.7037e-02,  9.3209e-02, -6.7559e-03,  7.1169e-02, -6.3749e-02,
    486.           6.1392e-02, -1.0353e-01, -4.0814e-02, -3.5229e-02,  1.0423e-01,
    487.          -7.9815e-02, -5.2447e-02, -7.7391e-02,  5.2047e-02,  2.9514e-02,
    488.           1.0629e-01, -1.0383e-02, -3.8161e-03, -3.2553e-02, -1.0170e-01,
    489.          -7.7791e-02, -6.3914e-02, -1.0758e-01,  2.8215e-02,  6.1443e-02,
    490.          -9.8031e-03, -8.3593e-02, -5.4196e-02, -5.4951e-02,  6.9776e-02,
    491.          -1.7996e-02, -3.2836e-02, -1.8449e-02,  3.7640e-02,  8.0527e-02,
    492.          -7.2464e-02,  4.5114e-02, -2.6867e-02, -1.0810e-01,  3.0985e-02,
    493.           8.5338e-02, -4.1375e-02, -7.6776e-02, -6.4114e-02, -5.0399e-02,
    494.           7.0755e-02,  6.3603e-02, -4.3264e-02, -2.7574e-03, -5.0588e-02,
    495.           4.6590e-02,  1.1015e-02, -4.8268e-02, -8.6557e-02,  4.6110e-02,
    496.           7.3038e-02,  8.7025e-02,  3.0157e-02,  2.0813e-02, -1.0674e-01,
    497.          -8.6806e-02, -3.2242e-02,  5.9878e-02,  7.4494e-02,  5.5173e-02,
    498.          -3.2275e-02, -2.4424e-02, -9.2330e-02,  4.2207e-02, -2.2701e-02,
    499.          -6.2829e-02,  1.0378e-01, -7.4888e-02, -5.0694e-02]],
    500.        requires_grad=True)
    501. name: fc3.bias, param: Parameter containing:
    502. tensor([ 0.0779,  0.1091, -0.0978, -0.0622, -0.0190, -0.0451,  0.0857,  0.0492,
    503.         -0.0665, -0.1034], requires_grad=True)
    504. output: tensor([[5.0619e+20, 9.3257e+20, 2.6026e+20, 0.0000e+00, 0.0000e+00, 5.8983e+20,
    505.          0.0000e+00, 2.4450e+20, 0.0000e+00, 2.6660e+20],
    506.         [4.6789e+20, 6.4717e+20, 9.8985e+19, 0.0000e+00, 0.0000e+00, 4.8509e+20,
    507.          0.0000e+00, 1.3280e+20, 0.0000e+00, 3.4584e+20],
    508.         [1.8287e+20, 7.6012e+20, 1.9911e+20, 0.0000e+00, 0.0000e+00, 4.5490e+20,
    509.          0.0000e+00, 0.0000e+00, 0.0000e+00, 1.9557e+20],
    510.         [       nan,        nan,        nan,        nan,        nan,        nan,
    511.                 nan,        nan,        nan,        nan],
    512.         [6.3816e+20, 8.7499e+20, 1.5305e+20, 0.0000e+00, 0.0000e+00, 4.0692e+20,
    513.          4.9000e+19, 5.2489e+19, 0.0000e+00, 2.9970e+20],
    514.         [3.2048e+20, 1.0051e+21, 0.0000e+00, 0.0000e+00, 0.0000e+00, 7.2725e+20,
    515.          0.0000e+00, 1.7769e+20, 0.0000e+00, 4.5969e+20],
    516.         [3.4913e+20, 1.0960e+21, 0.0000e+00, 0.0000e+00, 0.0000e+00, 4.3856e+20,
    517.          0.0000e+00, 4.9032e+19, 0.0000e+00, 2.6482e+20],
    518.         [3.7178e+29, 0.0000e+00, 2.3828e+29, 0.0000e+00, 0.0000e+00, 1.1230e+29,
    519.          0.0000e+00, 0.0000e+00, 1.7660e+28, 0.0000e+00],
    520.         [       nan,        nan,        nan,        nan,        nan,        nan,
    521.                 nan,        nan,        nan,        nan],
    522.         [       nan,        nan,        nan,        nan,        nan,        nan,
    523.                 nan,        nan,        nan,        nan]],
    524.        grad_fn=<ReluBackward0>)
    525. output.size: torch.Size([10, 10])