softmax实现cifar10分类

 将cifar10改成单一通道后，套用前面的softmax分类，分类率40%左右，想哭。。。

In [1]:

%matplotlib inline
from mxnet.gluon import data as gdata
from mxnet import autograd,nd
import gluonbook as gb
import sys

In [2]:

cifar_train = gdata.vision.CIFAR10(train=True)
cifar_test = gdata.vision.CIFAR10(train=False)

In [3]:

(len(cifar_train),len(cifar_test))

Out[3]:

(50000, 10000)

In [4]:

feature,label = cifar_train[0]

In [5]:

feature.shape,feature.dtype

Out[5]:

((32, 32, 3), numpy.uint8)

In [6]:

label,type(label),label.dtype

Out[6]:

(6, numpy.int32, dtype('int32'))

In [7]:

batch_size = 256
transformer = gdata.vision.transforms.ToTensor()

In [8]:

if sys.platform.startswith('win'):
    num_workers = 0  # 0 表示不用额外的进程来加速读取数据。
else:
    num_workers = 4

train_iter = gdata.DataLoader(cifar_train.transform_first(transformer),
                              batch_size, shuffle=True,
                              num_workers=num_workers)
test_iter = gdata.DataLoader(cifar_test.transform_first(transformer),
                             batch_size, shuffle=False,
                             num_workers=num_workers)

In [9]:

len(train_iter)

Out[9]:

196

In [10]:

for X,y in train_iter:
    print(X)
    break

 

[[[[0.3137255  0.3019608  0.34509805 ... 0.2901961  0.3019608
    0.34901962]
   [0.36078432 0.35686275 0.32941177 ... 0.23137255 0.2509804
    0.3764706 ]
   [0.34509805 0.42352942 0.47058824 ... 0.1882353  0.19607843
    0.3254902 ]
   ...
   [0.7529412  0.654902   0.5882353  ... 0.67058825 0.6627451
    0.78039217]
   [0.72156864 0.60784316 0.5764706  ... 0.63529414 0.63529414
    0.7372549 ]
   [0.65882355 0.6117647  0.6039216  ... 0.67058825 0.6627451
    0.6901961 ]]

  [[0.3137255  0.28627452 0.3137255  ... 0.28627452 0.29803923
    0.34509805]
   [0.36078432 0.34117648 0.3019608  ... 0.22745098 0.24705882
    0.37254903]
   [0.34509805 0.40392157 0.44313726 ... 0.18431373 0.19215687
    0.32156864]
   ...
   [0.8039216  0.7058824  0.6431373  ... 0.7019608  0.69803923
    0.8156863 ]
   [0.7764706  0.6627451  0.6313726  ... 0.6666667  0.6666667
    0.7764706 ]
   [0.7176471  0.6666667  0.65882355 ... 0.7019608  0.69803923
    0.7254902 ]]

  [[0.21960784 0.2        0.23137255 ... 0.21176471 0.21960784
    0.26666668]
   [0.26666668 0.2509804  0.21960784 ... 0.14901961 0.16862746
    0.29411766]
   [0.2509804  0.31764707 0.36078432 ... 0.10588235 0.11372549
    0.24313726]
   ...
   [0.6039216  0.5058824  0.4392157  ... 0.49803922 0.48235294
    0.5882353 ]
   [0.5764706  0.4627451  0.43137255 ... 0.46666667 0.4627451
    0.5529412 ]
   [0.5137255  0.46666667 0.45882353 ... 0.5137255  0.49803922
    0.5137255 ]]]


 [[[0.14901961 0.14901961 0.15294118 ... 0.14509805 0.09411765
    0.23137255]
   [0.15686275 0.15686275 0.16078432 ... 0.15686275 0.11372549
    0.2509804 ]
   [0.16078432 0.16470589 0.16862746 ... 0.16862746 0.12941177
    0.2627451 ]
   ...
   [0.16862746 0.12156863 0.14901961 ... 0.30588236 0.42352942
    0.24313726]
   [0.16862746 0.1254902  0.13333334 ... 0.28235295 0.39607844
    0.22352941]
   [0.16470589 0.1254902  0.09411765 ... 0.19607843 0.29411766
    0.16862746]]

  [[0.15294118 0.15294118 0.15686275 ... 0.15294118 0.09803922
    0.23529412]
   [0.16078432 0.16078432 0.16470589 ... 0.16470589 0.11764706
    0.25490198]
   [0.16470589 0.16862746 0.17254902 ... 0.1764706  0.13725491
    0.27058825]
   ...
   [0.17254902 0.1254902  0.14901961 ... 0.23137255 0.3019608
    0.19607843]
   [0.16862746 0.1254902  0.13333334 ... 0.22745098 0.28627452
    0.18039216]
   [0.16862746 0.12941177 0.09411765 ... 0.1764706  0.24705882
    0.14901961]]

  [[0.13333334 0.13333334 0.13725491 ... 0.15686275 0.09019608
    0.21568628]
   [0.14117648 0.14117648 0.14509805 ... 0.16862746 0.10980392
    0.23529412]
   [0.14509805 0.14901961 0.15294118 ... 0.18039216 0.1254902
    0.24705882]
   ...
   [0.14901961 0.10980392 0.13333334 ... 0.17254902 0.21960784
    0.15686275]
   [0.14901961 0.11372549 0.12156863 ... 0.18431373 0.20392157
    0.13333334]
   [0.14901961 0.11372549 0.08627451 ... 0.16078432 0.21176471
    0.1254902 ]]]


 [[[0.07843138 0.08627451 0.10196079 ... 0.0627451  0.05490196
    0.04705882]
   [0.10980392 0.08627451 0.11764706 ... 0.06666667 0.05490196
    0.04705882]
   [0.09019608 0.07058824 0.09411765 ... 0.05882353 0.05882353
    0.04705882]
   ...
   [0.18039216 0.16862746 0.1882353  ... 0.13725491 0.13725491
    0.13333334]
   [0.14901961 0.15294118 0.16470589 ... 0.14901961 0.12941177
    0.12156863]
   [0.13725491 0.14117648 0.15686275 ... 0.13725491 0.12156863
    0.11764706]]

  [[0.08627451 0.09411765 0.10980392 ... 0.07058824 0.0627451
    0.05490196]
   [0.12156863 0.09411765 0.1254902  ... 0.07450981 0.0627451
    0.05490196]
   [0.10588235 0.08235294 0.10196079 ... 0.06666667 0.06666667
    0.05490196]
   ...
   [0.19607843 0.1882353  0.2        ... 0.15294118 0.15294118
    0.14509805]
   [0.16470589 0.17254902 0.1764706  ... 0.16078432 0.14117648
    0.13333334]
   [0.15294118 0.16078432 0.16862746 ... 0.14901961 0.13333334
    0.12941177]]

  [[0.07058824 0.07843138 0.09019608 ... 0.05882353 0.05098039
    0.05098039]
   [0.10980392 0.07450981 0.10588235 ... 0.0627451  0.05490196
    0.05098039]
   [0.08627451 0.05882353 0.08627451 ... 0.05490196 0.05490196
    0.04705882]
   ...
   [0.16078432 0.14901961 0.16862746 ... 0.1254902  0.1254902
    0.12156863]
   [0.12941177 0.13333334 0.14117648 ... 0.13333334 0.11372549
    0.10588235]
   [0.11764706 0.1254902  0.13333334 ... 0.12156863 0.10588235
    0.10196079]]]


 ...


 [[[0.20784314 0.36078432 0.85490197 ... 0.972549   0.9647059
    0.96862745]
   [0.22745098 0.35686275 0.827451   ... 0.9764706  0.96862745
    0.9647059 ]
   [0.3372549  0.5019608  0.90588236 ... 0.9764706  0.9764706
    0.9647059 ]
   ...
   [0.08627451 0.08627451 0.05098039 ... 0.15294118 0.10980392
    0.09803922]
   [0.14901961 0.09411765 0.05098039 ... 0.10980392 0.18431373
    0.2784314 ]
   [0.3882353  0.27058825 0.14117648 ... 0.07058824 0.11764706
    0.16470589]]

  [[0.09803922 0.24705882 0.8156863  ... 0.9411765  0.9254902
    0.91764706]
   [0.14509805 0.25882354 0.7882353  ... 0.9372549  0.9254902
    0.8980392 ]
   [0.2784314  0.43137255 0.88235295 ... 0.9372549  0.9411765
    0.92941177]
   ...
   [0.06666667 0.07450981 0.05098039 ... 0.13725491 0.09411765
    0.08235294]
   [0.14117648 0.09019608 0.05098039 ... 0.09803922 0.17254902
    0.26666668]
   [0.3882353  0.27450982 0.14117648 ... 0.0627451  0.10980392
    0.15686275]]

  [[0.10588235 0.26666668 0.827451   ... 0.9607843  0.9411765
    0.92156863]
   [0.14117648 0.28627452 0.8156863  ... 0.94509804 0.9411765
    0.9254902 ]
   [0.27450982 0.4392157  0.88235295 ... 0.9254902  0.9490196
    0.96862745]
   ...
   [0.0627451  0.07058824 0.04313726 ... 0.13725491 0.09803922
    0.09019608]
   [0.13333334 0.08235294 0.04313726 ... 0.09803922 0.1764706
    0.27058825]
   [0.38039216 0.2627451  0.13333334 ... 0.06666667 0.11372549
    0.16078432]]]


 [[[0.35686275 0.33333334 0.34901962 ... 0.19607843 0.1882353
    0.1882353 ]
   [0.38431373 0.37254903 0.39215687 ... 0.25882354 0.27450982
    0.2627451 ]
   [0.38431373 0.38039216 0.3882353  ... 0.2509804  0.25490198
    0.24705882]
   ...
   [0.7764706  0.76862746 0.72156864 ... 0.76862746 0.77254903
    0.77254903]
   [0.77254903 0.7647059  0.77254903 ... 0.76862746 0.76862746
    0.77254903]
   [0.7647059  0.75686276 0.7529412  ... 0.75686276 0.7529412
    0.75686276]]

  [[0.35686275 0.3372549  0.34509805 ... 0.20784314 0.20392157
    0.19607843]
   [0.3882353  0.38039216 0.39607844 ... 0.26666668 0.2901961
    0.2627451 ]
   [0.3882353  0.38039216 0.3882353  ... 0.2509804  0.26666668
    0.25490198]
   ...
   [0.78039217 0.77254903 0.73333335 ... 0.76862746 0.77254903
    0.77254903]
   [0.77254903 0.7647059  0.77254903 ... 0.76862746 0.76862746
    0.77254903]
   [0.7647059  0.75686276 0.75686276 ... 0.7490196  0.7529412
    0.75686276]]

  [[0.2901961  0.2627451  0.28235295 ... 0.13725491 0.13725491
    0.13725491]
   [0.34901962 0.3372549  0.36078432 ... 0.20392157 0.21960784
    0.2       ]
   [0.36078432 0.3529412  0.37254903 ... 0.20784314 0.21568628
    0.21176471]
   ...
   [0.77254903 0.7607843  0.72156864 ... 0.7607843  0.7647059
    0.7647059 ]
   [0.7647059  0.75686276 0.7607843  ... 0.7607843  0.7607843
    0.7647059 ]
   [0.7607843  0.7529412  0.7490196  ... 0.74509805 0.74509805
    0.7490196 ]]]


 [[[0.8745098  0.8784314  0.8784314  ... 0.8235294  0.8
    0.7490196 ]
   [0.83137256 0.8235294  0.827451   ... 0.7647059  0.74509805
    0.73333335]
   [0.8039216  0.79607844 0.8039216  ... 0.67058825 0.6313726
    0.70980394]
   ...
   [0.40784314 0.3647059  0.34901962 ... 0.29803923 0.27450982
    0.28235295]
   [0.41568628 0.36078432 0.35686275 ... 0.26666668 0.25882354
    0.28627452]
   [0.3882353  0.3529412  0.34117648 ... 0.2784314  0.26666668
    0.28235295]]

  [[0.8901961  0.89411765 0.89411765 ... 0.8117647  0.8039216
    0.76862746]
   [0.84705883 0.8392157  0.84313726 ... 0.75686276 0.74509805
    0.7529412 ]
   [0.81960785 0.8117647  0.81960785 ... 0.6627451  0.6313726
    0.7294118 ]
   ...
   [0.3372549  0.31764707 0.30588236 ... 0.2784314  0.25490198
    0.2627451 ]
   [0.32156864 0.29803923 0.29411766 ... 0.23921569 0.23529412
    0.25882354]
   [0.29411766 0.28235295 0.27450982 ... 0.2509804  0.24705882
    0.25882354]]

  [[0.9372549  0.9411765  0.9411765  ... 0.85490197 0.8627451
    0.8352941 ]
   [0.89411765 0.8862745  0.8901961  ... 0.79607844 0.8039216
    0.81960785]
   [0.8666667  0.85882354 0.8666667  ... 0.7019608  0.6901961
    0.79607844]
   ...
   [0.23921569 0.20784314 0.19607843 ... 0.30588236 0.2627451
    0.2627451 ]
   [0.23529412 0.2        0.19607843 ... 0.26666668 0.23137255
    0.2509804 ]
   [0.21960784 0.2        0.1882353  ... 0.27058825 0.23921569
    0.2509804 ]]]]
<NDArray 256x3x32x32 @cpu(0)>

 

In [11]:

def wrapped_iter(data_iter):
    for X, y in data_iter:
        X = X[:, :1, :, :]
        yield X, y

for X, y in wrapped_iter(train_iter):
    print(X)
    print(y)
    break

for X, y in wrapped_iter(test_iter):
    print(X)
    print(y)
    break

 

[[[[0.40784314 0.3882353  0.40392157 ... 0.2509804  0.23921569
    0.22745098]
   [0.4        0.3882353  0.4        ... 0.2627451  0.2627451
    0.23529412]
   [0.39607844 0.38039216 0.4        ... 0.2901961  0.2901961
    0.26666668]
   ...
   [0.79607844 0.7882353  0.7882353  ... 0.59607846 0.58431375
    0.5764706 ]
   [0.74509805 0.7607843  0.74509805 ... 0.6431373  0.62352943
    0.6117647 ]
   [0.73333335 0.7254902  0.7372549  ... 0.6392157  0.6431373
    0.6313726 ]]]


 [[[1.         0.99215686 0.96862745 ... 0.62352943 0.6862745
    0.8627451 ]
   [1.         0.96862745 0.92156863 ... 0.5764706  0.6901961
    0.7607843 ]
   [1.         0.95686275 0.8745098  ... 0.63529414 0.7529412
    0.7607843 ]
   ...
   [0.49411765 0.5058824  0.58431375 ... 0.7019608  0.7294118
    0.7490196 ]
   [0.6431373  0.69803923 0.7254902  ... 0.7019608  0.7137255
    0.7176471 ]
   [0.8666667  0.9137255  0.8039216  ... 0.7058824  0.75686276
    0.77254903]]]


 [[[0.5411765  0.5411765  0.5647059  ... 0.29411766 0.21960784
    0.25882354]
   [0.58431375 0.56078434 0.5803922  ... 0.25490198 0.20392157
    0.26666668]
   [0.61960787 0.5686275  0.57254905 ... 0.23137255 0.21960784
    0.25882354]
   ...
   [0.59607846 0.6745098  0.70980394 ... 0.8352941  0.81960785
    0.8       ]
   [0.60784316 0.6901961  0.70980394 ... 0.8980392  0.91764706
    0.8156863 ]
   [0.6745098  0.75686276 0.7372549  ... 0.89411765 0.92156863
    0.9098039 ]]]


 ...


 [[[0.20392157 0.21176471 0.2        ... 0.14509805 0.16862746
    0.13725491]
   [0.19215687 0.20392157 0.21568628 ... 0.15294118 0.12156863
    0.09019608]
   [0.22352941 0.20784314 0.19607843 ... 0.21176471 0.17254902
    0.09803922]
   ...
   [0.49019608 0.47058824 0.5058824  ... 0.17254902 0.09411765
    0.14509805]
   [0.5019608  0.5882353  0.7019608  ... 0.1882353  0.18039216
    0.18039216]
   [0.42352942 0.5529412  0.68235296 ... 0.2        0.20784314
    0.23137255]]]


 [[[0.6431373  0.5803922  0.5921569  ... 0.24313726 0.3647059
    0.27450982]
   [0.69803923 0.6901961  0.5372549  ... 0.40392157 0.36078432
    0.2901961 ]
   [0.44705883 0.65882355 0.6        ... 0.49803922 0.3529412
    0.29411766]
   ...
   [0.827451   0.8039216  0.72156864 ... 0.25490198 0.25490198
    0.29411766]
   [0.89411765 0.8156863  0.7490196  ... 0.23529412 0.25882354
    0.2901961 ]
   [0.91764706 0.8392157  0.65882355 ... 0.22352941 0.22745098
    0.27058825]]]


 [[[0.04313726 0.07843138 0.14117648 ... 0.31764707 0.3254902
    0.25882354]
   [0.03529412 0.0627451  0.10980392 ... 0.3254902  0.28235295
    0.2627451 ]
   [0.01960784 0.05098039 0.07843138 ... 0.27450982 0.23529412
    0.2901961 ]
   ...
   [0.2627451  0.2901961  0.2509804  ... 0.32941177 0.34901962
    0.3254902 ]
   [0.24313726 0.21176471 0.1882353  ... 0.32941177 0.3137255
    0.28627452]
   [0.28235295 0.24705882 0.21960784 ... 0.3254902  0.29411766
    0.26666668]]]]
<NDArray 256x1x32x32 @cpu(0)>

[2 9 4 7 3 1 3 5 9 6 2 9 4 4 9 5 3 7 2 9 3 2 1 4 3 1 0 6 7 4 4 0 5 6 3 3 8
 2 6 1 8 1 4 0 7 1 4 8 4 5 1 0 6 8 1 0 8 4 4 7 0 9 9 2 6 4 4 2 7 3 4 3 0 0
 9 2 4 0 7 6 5 9 6 5 0 0 0 6 7 8 8 7 7 8 7 9 3 4 4 6 1 0 5 6 0 6 6 7 1 8 9
 2 2 5 2 9 9 8 6 2 4 3 1 7 0 2 4 8 3 6 3 7 2 4 4 9 2 3 7 0 6 9 4 9 6 6 7 6
 8 2 5 4 7 6 0 2 9 5 9 3 1 5 9 2 1 7 7 0 5 0 5 2 3 9 7 1 3 5 5 7 0 6 2 3 1
 5 3 6 2 2 5 7 0 7 5 8 5 9 7 0 7 2 8 1 7 4 2 3 8 6 1 6 1 6 0 8 8 8 7 9 4 2
 6 6 9 1 5 2 5 1 4 6 1 8 9 2 4 7 0 4 3 3 6 5 9 4 1 0 2 5 9 3 1 6 6 6]
<NDArray 256 @cpu(0)>

[[[[0.61960787 0.62352943 0.64705884 ... 0.5372549  0.49411765
    0.45490196]
   [0.59607846 0.5921569  0.62352943 ... 0.53333336 0.49019608
    0.46666667]
   [0.5921569  0.5921569  0.61960787 ... 0.54509807 0.50980395
    0.47058824]
   ...
   [0.26666668 0.16470589 0.12156863 ... 0.14901961 0.05098039
    0.15686275]
   [0.23921569 0.19215687 0.13725491 ... 0.10196079 0.11372549
    0.07843138]
   [0.21176471 0.21960784 0.1764706  ... 0.09411765 0.13333334
    0.08235294]]]


 [[[0.92156863 0.90588236 0.9098039  ... 0.9137255  0.9137255
    0.9098039 ]
   [0.93333334 0.92156863 0.92156863 ... 0.9254902  0.9254902
    0.92156863]
   [0.92941177 0.91764706 0.91764706 ... 0.92156863 0.92156863
    0.91764706]
   ...
   [0.34117648 0.16862746 0.07450981 ... 0.6627451  0.7137255
    0.7372549 ]
   [0.32156864 0.18039216 0.14117648 ... 0.68235296 0.7254902
    0.73333335]
   [0.33333334 0.24313726 0.22745098 ... 0.65882355 0.7058824
    0.7294118 ]]]


 [[[0.61960787 0.61960787 0.54509807 ... 0.89411765 0.92941177
    0.93333334]
   [0.6666667  0.6745098  0.5921569  ... 0.9098039  0.9647059
    0.9647059 ]
   [0.68235296 0.6901961  0.6156863  ... 0.9019608  0.98039216
    0.9607843 ]
   ...
   [0.12156863 0.11764706 0.10196079 ... 0.14509805 0.03529412
    0.01568628]
   [0.09019608 0.10588235 0.09803922 ... 0.07450981 0.01568628
    0.01960784]
   [0.10980392 0.11764706 0.1254902  ... 0.01960784 0.01568628
    0.02745098]]]


 ...


 [[[0.2627451  0.26666668 0.27450982 ... 0.28235295 0.2784314
    0.27450982]
   [0.27058825 0.2784314  0.28627452 ... 0.2901961  0.2901961
    0.28627452]
   [0.2784314  0.28235295 0.28627452 ... 0.29411766 0.2901961
    0.28627452]
   ...
   [0.35686275 0.3882353  0.37254903 ... 0.30980393 0.34901962
    0.3647059 ]
   [0.33333334 0.35686275 0.34901962 ... 0.27058825 0.26666668
    0.28235295]
   [0.3254902  0.3372549  0.33333334 ... 0.2627451  0.26666668
    0.25882354]]]


 [[[0.7254902  0.7058824  0.6745098  ... 0.6156863  0.59607846
    0.54901963]
   [0.7921569  0.69411767 0.63529414 ... 0.6039216  0.5764706
    0.5529412 ]
   [0.7176471  0.6392157  0.627451   ... 0.5764706  0.5764706
    0.5803922 ]
   ...
   [0.6901961  0.62352943 0.6156863  ... 0.37254903 0.31764707
    0.29803923]
   [0.6784314  0.6392157  0.67058825 ... 0.39215687 0.38431373
    0.36078432]
   [0.64705884 0.59607846 0.62352943 ... 0.47843137 0.5176471
    0.46666667]]]


 [[[0.8        0.8039216  0.8156863  ... 0.8352941  0.84705883
    0.84705883]
   [0.80784315 0.8156863  0.827451   ... 0.8352941  0.8235294
    0.827451  ]
   [0.7882353  0.7921569  0.80784315 ... 0.78431374 0.76862746
    0.76862746]
   ...
   [0.5058824  0.50980395 0.52156866 ... 0.45882353 0.5137255
    0.5294118 ]
   [0.49411765 0.49803922 0.5058824  ... 0.4627451  0.5176471
    0.5254902 ]
   [0.4862745  0.49019608 0.49803922 ... 0.4509804  0.49803922
    0.5058824 ]]]]
<NDArray 256x1x32x32 @cpu(0)>

[3 8 8 0 6 6 1 6 3 1 0 9 5 7 9 8 5 7 8 6 7 0 4 9 5 2 4 0 9 6 6 5 4 5 9 2 4
 1 9 5 4 6 5 6 0 9 3 9 7 6 9 8 0 3 8 8 7 7 4 6 7 3 6 3 6 2 1 2 3 7 2 6 8 8
 0 2 9 3 3 8 8 1 1 7 2 5 2 7 8 9 0 3 8 6 4 6 6 0 0 7 4 5 6 3 1 1 3 6 8 7 4
 0 6 2 1 3 0 4 2 7 8 3 1 2 8 0 8 3 5 2 4 1 8 9 1 2 9 7 2 9 6 5 6 3 8 7 6 2
 5 2 8 9 6 0 0 5 2 9 5 4 2 1 6 6 8 4 8 4 5 0 9 9 9 8 9 9 3 7 5 0 0 5 2 2 3
 8 6 3 4 0 5 8 0 1 7 2 8 8 7 8 5 1 8 7 1 3 0 5 7 9 7 4 5 9 8 0 7 9 8 2 7 6
 9 4 3 9 6 4 7 6 5 1 5 8 8 0 4 0 5 5 1 1 8 9 0 3 1 9 2 2 5 3 9 9 4 0]
<NDArray 256 @cpu(0)>

In [12]:

from mxnet import gluon, init
from mxnet.gluon import loss as gloss, nn

 

In [13]:

net = nn.Sequential()
net.add(nn.Dense(10))
net.initialize(init.Normal(sigma=0.01))

In [14]:

loss = gloss.SoftmaxCrossEntropyLoss()

In [25]:

trainer = gluon.Trainer(net.collect_params(), 'sgd', {'learning_rate': 0.0001})

In [26]:

num_epochs = 100
gb.train_ch3(net, train_iter, test_iter, loss, num_epochs, batch_size, None,
             None, trainer)

 

epoch 1, loss 1.6195, train acc 0.457, test acc 0.410
epoch 2, loss 1.6196, train acc 0.457, test acc 0.411
epoch 3, loss 1.6181, train acc 0.457, test acc 0.411
epoch 4, loss 1.6183, train acc 0.457, test acc 0.411
epoch 5, loss 1.6191, train acc 0.457, test acc 0.410
epoch 6, loss 1.6196, train acc 0.457, test acc 0.411
epoch 7, loss 1.6189, train acc 0.457, test acc 0.410
epoch 8, loss 1.6189, train acc 0.457, test acc 0.411
epoch 9, loss 1.6183, train acc 0.457, test acc 0.410
epoch 10, loss 1.6186, train acc 0.457, test acc 0.411
epoch 11, loss 1.6182, train acc 0.457, test acc 0.410
epoch 12, loss 1.6175, train acc 0.457, test acc 0.410
epoch 13, loss 1.6181, train acc 0.457, test acc 0.410
epoch 14, loss 1.6182, train acc 0.457, test acc 0.411
epoch 15, loss 1.6192, train acc 0.457, test acc 0.410
epoch 16, loss 1.6191, train acc 0.457, test acc 0.411
epoch 17, loss 1.6182, train acc 0.457, test acc 0.410
epoch 18, loss 1.6176, train acc 0.457, test acc 0.410
epoch 19, loss 1.6175, train acc 0.458, test acc 0.410
epoch 20, loss 1.6182, train acc 0.457, test acc 0.410
epoch 21, loss 1.6178, train acc 0.457, test acc 0.410
epoch 22, loss 1.6180, train acc 0.457, test acc 0.410
epoch 23, loss 1.6178, train acc 0.457, test acc 0.411
epoch 24, loss 1.6179, train acc 0.457, test acc 0.411
epoch 25, loss 1.6178, train acc 0.457, test acc 0.411
epoch 26, loss 1.6180, train acc 0.457, test acc 0.411
epoch 27, loss 1.6181, train acc 0.457, test acc 0.410
epoch 28, loss 1.6172, train acc 0.457, test acc 0.410
epoch 29, loss 1.6177, train acc 0.457, test acc 0.411
epoch 30, loss 1.6170, train acc 0.458, test acc 0.410
epoch 31, loss 1.6162, train acc 0.458, test acc 0.410
epoch 32, loss 1.6184, train acc 0.457, test acc 0.410
epoch 33, loss 1.6175, train acc 0.457, test acc 0.410
epoch 34, loss 1.6174, train acc 0.457, test acc 0.411
epoch 35, loss 1.6173, train acc 0.457, test acc 0.411
epoch 36, loss 1.6177, train acc 0.457, test acc 0.411
epoch 37, loss 1.6174, train acc 0.457, test acc 0.410
epoch 38, loss 1.6174, train acc 0.457, test acc 0.410
epoch 39, loss 1.6171, train acc 0.457, test acc 0.411
epoch 40, loss 1.6178, train acc 0.457, test acc 0.410
epoch 41, loss 1.6173, train acc 0.457, test acc 0.410
epoch 42, loss 1.6169, train acc 0.457, test acc 0.411
epoch 43, loss 1.6166, train acc 0.457, test acc 0.410
epoch 44, loss 1.6172, train acc 0.457, test acc 0.410
epoch 45, loss 1.6166, train acc 0.457, test acc 0.410
epoch 46, loss 1.6174, train acc 0.457, test acc 0.410
epoch 47, loss 1.6170, train acc 0.457, test acc 0.410
epoch 48, loss 1.6166, train acc 0.457, test acc 0.410
epoch 49, loss 1.6165, train acc 0.457, test acc 0.410
epoch 50, loss 1.6163, train acc 0.457, test acc 0.410
epoch 51, loss 1.6167, train acc 0.457, test acc 0.410
epoch 52, loss 1.6172, train acc 0.457, test acc 0.410
epoch 53, loss 1.6163, train acc 0.458, test acc 0.410
epoch 54, loss 1.6166, train acc 0.457, test acc 0.410
epoch 55, loss 1.6163, train acc 0.457, test acc 0.410
epoch 56, loss 1.6171, train acc 0.457, test acc 0.410
epoch 57, loss 1.6170, train acc 0.457, test acc 0.410
epoch 58, loss 1.6163, train acc 0.457, test acc 0.410
epoch 59, loss 1.6160, train acc 0.458, test acc 0.410
epoch 60, loss 1.6163, train acc 0.457, test acc 0.410
epoch 61, loss 1.6165, train acc 0.457, test acc 0.410
epoch 62, loss 1.6157, train acc 0.457, test acc 0.410
epoch 63, loss 1.6169, train acc 0.457, test acc 0.410
epoch 64, loss 1.6158, train acc 0.457, test acc 0.410
epoch 65, loss 1.6167, train acc 0.457, test acc 0.410
epoch 66, loss 1.6162, train acc 0.458, test acc 0.410
epoch 67, loss 1.6167, train acc 0.457, test acc 0.410
epoch 68, loss 1.6163, train acc 0.457, test acc 0.409
epoch 69, loss 1.6170, train acc 0.457, test acc 0.410
epoch 70, loss 1.6164, train acc 0.457, test acc 0.410
epoch 71, loss 1.6166, train acc 0.457, test acc 0.410
epoch 72, loss 1.6157, train acc 0.457, test acc 0.410
epoch 73, loss 1.6159, train acc 0.457, test acc 0.410
epoch 74, loss 1.6163, train acc 0.457, test acc 0.410
epoch 75, loss 1.6162, train acc 0.457, test acc 0.410
epoch 76, loss 1.6154, train acc 0.457, test acc 0.409
epoch 77, loss 1.6161, train acc 0.457, test acc 0.410
epoch 78, loss 1.6169, train acc 0.457, test acc 0.409
epoch 79, loss 1.6154, train acc 0.457, test acc 0.409
epoch 80, loss 1.6162, train acc 0.457, test acc 0.409
epoch 81, loss 1.6163, train acc 0.457, test acc 0.410
epoch 82, loss 1.6161, train acc 0.457, test acc 0.409
epoch 83, loss 1.6156, train acc 0.457, test acc 0.410
epoch 84, loss 1.6153, train acc 0.458, test acc 0.409
epoch 85, loss 1.6159, train acc 0.457, test acc 0.409
epoch 86, loss 1.6164, train acc 0.457, test acc 0.410
epoch 87, loss 1.6154, train acc 0.457, test acc 0.410
epoch 88, loss 1.6152, train acc 0.457, test acc 0.410
epoch 89, loss 1.6154, train acc 0.457, test acc 0.410
epoch 90, loss 1.6155, train acc 0.457, test acc 0.409
epoch 91, loss 1.6160, train acc 0.458, test acc 0.409
epoch 92, loss 1.6148, train acc 0.458, test acc 0.409
epoch 93, loss 1.6156, train acc 0.457, test acc 0.409
epoch 94, loss 1.6152, train acc 0.457, test acc 0.409
epoch 95, loss 1.6157, train acc 0.458, test acc 0.410
epoch 96, loss 1.6152, train acc 0.458, test acc 0.410
epoch 97, loss 1.6152, train acc 0.457, test acc 0.410
epoch 98, loss 1.6151, train acc 0.457, test acc 0.410
epoch 99, loss 1.6150, train acc 0.457, test acc 0.409
epoch 100, loss 1.6158, train acc 0.457, test acc 0.410

In [17]:

gb.train_ch3??

In [ ]: