前言

前文我们实现了googlenet。想必大家已经有感觉，实现一个神经网络不是很难嘛。但是实现一个神经网络知识深度学习的入门和开始，如何训练和使用一个网络将它使用和应用到工作和需求中，才是我们学习神经网络的初衷。所以今天我们就开始看看怎么使用前文实现的这个神经网络。

googlenet在pytorch里面已经有实现了，同时pytorch还提供了一个使用imagenet 1000分类训练过的权重。所以我们这里直接使用这个预训练模型，然后调整网络结构，将输出变为10分类，使用cifar10数据集重新训练（filetuning）一个属于我们自己的模型。pytorch的google可以从：https://pytorch.org/hub/pytorch_vision_googlenet/[1] 这里根据示例代码直接下载使用 也可以从https://github.com/pytorch/vision/blob/master/torchvision/models/googlenet.py [2]找到源代码。

加载预训练模型

我们这里使用[2] 提供的googlenet实现，[2]中同时提供了加载imagenet预训练权重的函数，我们进行执行对应的代码加载即可，代码如下：

import warnings
from collections import namedtuple
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch import Tensor
try:
    from torch.hub import load_state_dict_from_url
except ImportError:
    from torch.utils.model_zoo import load_url as load_state_dict_from_url
#from .utils import load_state_dict_from_url
from typing import Optional, Tuple, List, Callable, Any

__all__ = ['GoogLeNet', 'googlenet', "GoogLeNetOutputs", "_GoogLeNetOutputs"]

model_urls = {
    # GoogLeNet ported from TensorFlow
    'googlenet': 'https://download.pytorch.org/models/googlenet-1378be20.pth',
}

GoogLeNetOutputs = namedtuple('GoogLeNetOutputs', ['logits', 'aux_logits2', 'aux_logits1'])
GoogLeNetOutputs.__annotations__ = {'logits': Tensor, 'aux_logits2': Optional[Tensor],
                                    'aux_logits1': Optional[Tensor]}

# Script annotations failed with _GoogleNetOutputs = namedtuple ...
# _GoogLeNetOutputs set here for backwards compat
_GoogLeNetOutputs = GoogLeNetOutputs


def googlenet(pretrained: bool = False, progress: bool = True, **kwargs: Any) -> "GoogLeNet":
    r"""GoogLeNet (Inception v1) model architecture from
    `"Going Deeper with Convolutions" <http://arxiv.org/abs/1409.4842>`_.
    Args:
        pretrained (bool): If True, returns a model pre-trained on ImageNet
        progress (bool): If True, displays a progress bar of the download to stderr
        aux_logits (bool): If True, adds two auxiliary branches that can improve training.
            Default: *False* when pretrained is True otherwise *True*
        transform_input (bool): If True, preprocesses the input according to the method with which it
            was trained on ImageNet. Default: *False*
    """
    if pretrained:
        if 'transform_input' not in kwargs:
            kwargs['transform_input'] = True
        if 'aux_logits' not in kwargs:
            kwargs['aux_logits'] = False
        if kwargs['aux_logits']:
            warnings.warn('auxiliary heads in the pretrained googlenet model are NOT pretrained, '
                          'so make sure to train them')
        original_aux_logits = kwargs['aux_logits']
        kwargs['aux_logits'] = True
        kwargs['init_weights'] = False
        model = GoogLeNet(**kwargs)
        state_dict = load_state_dict_from_url(model_urls['googlenet'],
                                              progress=progress)
        model.load_state_dict(state_dict)
        if not original_aux_logits:
            model.aux_logits = False
            model.aux1 = None  # type: ignore[assignment]
            model.aux2 = None  # type: ignore[assignment]
        return model

    return GoogLeNet(**kwargs)


class GoogLeNet(nn.Module):
    __constants__ = ['aux_logits', 'transform_input']

    def __init__(
        self,
        num_classes: int = 1000,
        aux_logits: bool = True,
        transform_input: bool = False,
        init_weights: Optional[bool] = None,
        blocks: Optional[List[Callable[..., nn.Module]]] = None
    ) -> None:
        super(GoogLeNet, self).__init__()
        if blocks is None:
            blocks = [BasicConv2d, Inception, InceptionAux]
        if init_weights is None:
            warnings.warn('The default weight initialization of GoogleNet will be changed in future releases of '
                          'torchvision. If you wish to keep the old behavior (which leads to long initialization times'
                          ' due to scipy/scipy#11299), please set init_weights=True.', FutureWarning)
            init_weights = True
        assert len(blocks) == 3
        conv_block = blocks[0]
        inception_block = blocks[1]
        inception_aux_block = blocks[2]

        self.aux_logits = aux_logits
        self.transform_input = transform_input

        self.conv1 = conv_block(3, 64, kernel_size=7, stride=2, padding=3)
        self.maxpool1 = nn.MaxPool2d(3, stride=2, ceil_mode=True)
        self.conv2 = conv_block(64, 64, kernel_size=1)
        self.conv3 = conv_block(64, 192, kernel_size=3, padding=1)
        self.maxpool2 = nn.MaxPool2d(3, stride=2, ceil_mode=True)

        self.inception3a = inception_block(192, 64, 96, 128, 16, 32, 32)
        self.inception3b = inception_block(256, 128, 128, 192, 32, 96, 64)
        self.maxpool3 = nn.MaxPool2d(3, stride=2, ceil_mode=True)

        self.inception4a = inception_block(480, 192, 96, 208, 16, 48, 64)
        self.inception4b = inception_block(512, 160, 112, 224, 24, 64, 64)
        self.inception4c = inception_block(512, 128, 128, 256, 24, 64, 64)
        self.inception4d = inception_block(512, 112, 144, 288, 32, 64, 64)
        self.inception4e = inception_block(528, 256, 160, 320, 32, 128, 128)
        self.maxpool4 = nn.MaxPool2d(2, stride=2, ceil_mode=True)

        self.inception5a = inception_block(832, 256, 160, 320, 32, 128, 128)
        self.inception5b = inception_block(832, 384, 192, 384, 48, 128, 128)

        if aux_logits:
            self.aux1 = inception_aux_block(512, num_classes)
            self.aux2 = inception_aux_block(528, num_classes)
        else:
            self.aux1 = None  # type: ignore[assignment]
            self.aux2 = None  # type: ignore[assignment]

        self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
        self.dropout = nn.Dropout(0.2)
        self.fc = nn.Linear(1024, num_classes)
        
        

        if init_weights:
            self._initialize_weights()

    def _initialize_weights(self) -> None:
        for m in self.modules():
            if isinstance(m, nn.Conv2d) or isinstance(m, nn.Linear):
                import scipy.stats as stats
                X = stats.truncnorm(-2, 2, scale=0.01)
                values = torch.as_tensor(X.rvs(m.weight.numel()), dtype=m.weight.dtype)
                values = values.view(m.weight.size())
                with torch.no_grad():
                    m.weight.copy_(values)
            elif isinstance(m, nn.BatchNorm2d):
                nn.init.constant_(m.weight, 1)
                nn.init.constant_(m.bias, 0)

    def _transform_input(self, x: Tensor) -> Tensor:
        if self.transform_input:
            x_ch0 = torch.unsqueeze(x[:, 0], 1) * (0.229 / 0.5) + (0.485 - 0.5) / 0.5
            x_ch1 = torch.unsqueeze(x[:, 1], 1) * (0.224 / 0.5) + (0.456 - 0.5) / 0.5
            x_ch2 = torch.unsqueeze(x[:, 2], 1) * (0.225 / 0.5) + (0.406 - 0.5) / 0.5
            x = torch.cat((x_ch0, x_ch1, x_ch2), 1)
        return x

    def _forward(self, x: Tensor) -> Tuple[Tensor, Optional[Tensor], Optional[Tensor]]:
        # N x 3 x 224 x 224
        x = self.conv1(x)
        # N x 64 x 112 x 112
        x = self.maxpool1(x)
        # N x 64 x 56 x 56
        x = self.conv2(x)
        # N x 64 x 56 x 56
        x = self.conv3(x)
        # N x 192 x 56 x 56
        x = self.maxpool2(x)

        # N x 192 x 28 x 28
        x = self.inception3a(x)
        # N x 256 x 28 x 28
        x = self.inception3b(x)
        # N x 480 x 28 x 28
        x = self.maxpool3(x)
        # N x 480 x 14 x 14
        x = self.inception4a(x)
        # N x 512 x 14 x 14
        aux1 = torch.jit.annotate(Optional[Tensor], None)
        if self.aux1 is not None:
            if self.training:
                aux1 = self.aux1(x)

        x = self.inception4b(x)
        # N x 512 x 14 x 14
        x = self.inception4c(x)
        # N x 512 x 14 x 14
        x = self.inception4d(x)
        # N x 528 x 14 x 14
        aux2 = torch.jit.annotate(Optional[Tensor], None)
        if self.aux2 is not None:
            if self.training:
                aux2 = self.aux2(x)

        x = self.inception4e(x)
        # N x 832 x 14 x 14
        x = self.maxpool4(x)
        # N x 832 x 7 x 7
        x = self.inception5a(x)
        # N x 832 x 7 x 7
        x = self.inception5b(x)
        # N x 1024 x 7 x 7

        x = self.avgpool(x)
        # N x 1024 x 1 x 1
        x = torch.flatten(x, 1)
        # N x 1024
        x = self.dropout(x)
        x = self.fc(x)
        # N x 1000 (num_classes)
        return x, aux2, aux1

    @torch.jit.unused
    def eager_outputs(self, x: Tensor, aux2: Tensor, aux1: Optional[Tensor]) -> GoogLeNetOutputs:
        if self.training and self.aux_logits:
            return _GoogLeNetOutputs(x, aux2, aux1)
        else:
            return x   # type: ignore[return-value]

    def forward(self, x: Tensor) -> GoogLeNetOutputs:
        x = self._transform_input(x)
        x, aux1, aux2 = self._forward(x)
        aux_defined = self.training and self.aux_logits
        if torch.jit.is_scripting():
            if not aux_defined:
                warnings.warn("Scripted GoogleNet always returns GoogleNetOutputs Tuple")
            return GoogLeNetOutputs(x, aux2, aux1)
        else:
            return self.eager_outputs(x, aux2, aux1)


class Inception(nn.Module):

    def __init__(
        self,
        in_channels: int,
        ch1x1: int,
        ch3x3red: int,
        ch3x3: int,
        ch5x5red: int,
        ch5x5: int,
        pool_proj: int,
        conv_block: Optional[Callable[..., nn.Module]] = None
    ) -> None:
        super(Inception, self).__init__()
        if conv_block is None:
            conv_block = BasicConv2d
        self.branch1 = conv_block(in_channels, ch1x1, kernel_size=1)

        self.branch2 = nn.Sequential(
            conv_block(in_channels, ch3x3red, kernel_size=1),
            conv_block(ch3x3red, ch3x3, kernel_size=3, padding=1)
        )

        self.branch3 = nn.Sequential(
            conv_block(in_channels, ch5x5red, kernel_size=1),
            # Here, kernel_size=3 instead of kernel_size=5 is a known bug.
            # Please see https://github.com/pytorch/vision/issues/906 for details.
            conv_block(ch5x5red, ch5x5, kernel_size=3, padding=1)
        )

        self.branch4 = nn.Sequential(
            nn.MaxPool2d(kernel_size=3, stride=1, padding=1, ceil_mode=True),
            conv_block(in_channels, pool_proj, kernel_size=1)
        )

    def _forward(self, x: Tensor) -> List[Tensor]:
        branch1 = self.branch1(x)
        branch2 = self.branch2(x)
        branch3 = self.branch3(x)
        branch4 = self.branch4(x)

        outputs = [branch1, branch2, branch3, branch4]
        return outputs

    def forward(self, x: Tensor) -> Tensor:
        outputs = self._forward(x)
        return torch.cat(outputs, 1)


class InceptionAux(nn.Module):

    def __init__(
        self,
        in_channels: int,
        num_classes: int,
        conv_block: Optional[Callable[..., nn.Module]] = None
    ) -> None:
        super(InceptionAux, self).__init__()
        if conv_block is None:
            conv_block = BasicConv2d
        self.conv = conv_block(in_channels, 128, kernel_size=1)

        self.fc1 = nn.Linear(2048, 1024)
        self.fc2 = nn.Linear(1024, num_classes)

    def forward(self, x: Tensor) -> Tensor:
        # aux1: N x 512 x 14 x 14, aux2: N x 528 x 14 x 14
        x = F.adaptive_avg_pool2d(x, (4, 4))
        # aux1: N x 512 x 4 x 4, aux2: N x 528 x 4 x 4
        x = self.conv(x)
        # N x 128 x 4 x 4
        x = torch.flatten(x, 1)
        # N x 2048
        x = F.relu(self.fc1(x), inplace=True)
        # N x 1024
        x = F.dropout(x, 0.7, training=self.training)
        # N x 1024
        x = self.fc2(x)
        # N x 1000 (num_classes)

        return x


class BasicConv2d(nn.Module):

    def __init__(
        self,
        in_channels: int,
        out_channels: int,
        **kwargs: Any
    ) -> None:
        super(BasicConv2d, self).__init__()
        self.conv = nn.Conv2d(in_channels, out_channels, bias=False, **kwargs)
        self.bn = nn.BatchNorm2d(out_channels, eps=0.001)

    def forward(self, x: Tensor) -> Tensor:
        x = self.conv(x)
        x = self.bn(x)
        return F.relu(x, inplace=True)
        
model = googlenet(True)
p1 = next(model.parameters())
# change the fc layers to fit cifar10 dataset.
model.fc = nn.Linear(1024, 10)

注意到 原始代码中的 load_state_dict_from_url 是通过utils.py简洁import进来的，我们这里直接将其import当前代码中。另外对于InceptionAux模块以及对应的aux head，在进行finetuing的时候需要进行重新训练（pytorch提供的预训练权重未包含auxhead的权重，我们这里的实现里面，并没有finetuning aux head，aux head 和分类head的finetuning原理是一样的）。

我们使用googlenet函数加载预训练模型之后，直接替换model的fc层 （见行326）。这样模型

中其他层都是预训练的，可以用来做特征抽取使用（暂且称为特征抽取层）。我们的finetuning这里主要就是要修改更新fc这一层的权重（这里称为分类层或者分类head）。

值得注意的是，我们可以在使用自己的数据集重新训练的过程中，不更新特征抽取层权重。这种情况下，训练只会调整分类头。这对于数据集与pretrain数据集类似的情况（或者数据集很小不容易训练）比较合适。相反，我们可以选择训练过程中更新特征抽取层权重（极端情况就是完全重新训练），这对于数据集较大且和pretrain数据集区别较大的情况比较合适。

我们这里实现的是 更新特征抽取层的 finetuning（pytorch中， 要禁止某层的权重参数更新只需要将对应属性置为false即可： param.requires_grad = False）

加载cifar10数据

有了加载了pretrain权重并且修改了分类head的模型，我们这里准备下训练数据。这里我们选择cifar10 数据集：http://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz

这个数据集里面的数据解压完了是一些二进制数据，所以需要进行处理一下才能使用，如下代码：

如代码所示：

import pickle
import numpy as np
from torchvision import transforms
from PIL import Image

def unpickle(file):
    """load the cifar-10 data"""

    with open(file, 'rb') as fo:
        data = pickle.load(fo, encoding='bytes')
    return data

def load_cifar_10_data(data_dir, negatives=False):

    """
    Return train_data, train_filenames, train_labels, test_data, test_filenames, test_labels
    """

    # get the meta_data_dict
    # num_cases_per_batch: 1000
    # label_names: ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']
    # num_vis: :3072

    meta_data_dict = unpickle(data_dir + "/batches.meta")
    cifar_label_names = meta_data_dict[b'label_names']
    cifar_label_names = np.array(cifar_label_names)

    # training data
    cifar_train_data = None
    cifar_train_filenames = []
    cifar_train_labels = []

    # cifar_train_data_dict
    # 'batch_label': 'training batch 5 of 5'
    # 'data': ndarray
    # 'filenames': list
    # 'labels': list

    for i in range(1, 6):
        cifar_train_data_dict = unpickle(data_dir + "/data_batch_{}".format(i))
        if i == 1:
            cifar_train_data = cifar_train_data_dict[b'data']
        else:
            cifar_train_data = np.vstack((cifar_train_data, cifar_train_data_dict[b'data']))
        cifar_train_filenames += cifar_train_data_dict[b'filenames']
        cifar_train_labels += cifar_train_data_dict[b'labels']

    cifar_train_data = cifar_train_data.reshape((len(cifar_train_data), 3, 32, 32))
    if negatives:
        cifar_train_data = cifar_train_data.transpose(0, 2, 3, 1).astype(np.float32)
    else:
        cifar_train_data = np.rollaxis(cifar_train_data, 1, 4)
    cifar_train_filenames = np.array(cifar_train_filenames)
    cifar_train_labels = np.array(cifar_train_labels)

    # test data
    # cifar_test_data_dict
    # 'batch_label': 'testing batch 1 of 1'
    # 'data': ndarray
    # 'filenames': list
    # 'labels': list

    cifar_test_data_dict = unpickle(data_dir + "/test_batch")
    cifar_test_data = cifar_test_data_dict[b'data']
    cifar_test_filenames = cifar_test_data_dict[b'filenames']
    cifar_test_labels = cifar_test_data_dict[b'labels']

    cifar_test_data = cifar_test_data.reshape((len(cifar_test_data), 3, 32, 32))
    if negatives:
        cifar_test_data = cifar_test_data.transpose(0, 2, 3, 1).astype(np.float32)
    else:
        cifar_test_data = np.rollaxis(cifar_test_data, 1, 4)
    cifar_test_filenames = np.array(cifar_test_filenames)
    cifar_test_labels = np.array(cifar_test_labels)

    return cifar_train_data, cifar_train_filenames, cifar_train_labels, \
        cifar_test_data, cifar_test_filenames, cifar_test_labels, cifar_label_names
        
        train_data, train_filenames, train_labels, test_data, test_filenames, test_labels, label_names\
    = load_cifar_10_data(fpath)



def make_data_loader(data, labels):
    ret = {}
    train_factor = 0.8
    print(data.shape)
    print(type(data))
    train_num = int(data.shape[0]*train_factor)
    
    def makeTensor(da):
        preprocess = transforms.Compose([
            transforms.Resize(256),
            transforms.CenterCrop(224),
            transforms.ToTensor(),
            transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
        ])
        input_tensor = preprocess(Image.fromarray(da))
        input_batch = input_tensor.unsqueeze(0)
        return input_batch
    
    def make_one_hot(id):
        #arr = np.array([0 for i in range(10)])
        #arr[id] = 1
        return torch.from_numpy(np.array([id]))
        #pre_ret = torch.from_numpy(arr)
        #eturn pre_ret.unsqueeze(0)
    
    data = data[:100,:]
    labels = labels[:100]
    
    t_data = data[0:train_num, :]
    t_data = [t_data[idx,:] for idx in range(len(t_data))]
    
    t_label = labels[0:train_num]
    t_label = [t_label[idx] for idx in range(len(t_label))]
    
    ret_train = [(makeTensor(t_data[idx]),make_one_hot(t_label[idx])) for idx in range(len(t_data)) ]
    
    
    
    
    v_data = data[:train_num:,]
    v_label = labels[:train_num:,]
     
    v_data = [v_data[idx,:] for idx in range(len(v_data))]
    v_label = [v_label[idx] for idx in range(len(v_label))]
    ret_val = [ (makeTensor(v_data[idx]), make_one_hot(v_label[idx])) for idx in range(len(v_data))]
    
    ret["train"] = ret_train
    ret["val"] = ret_val
    
    return ret
    

loader = make_data_loader(train_data, train_labels)

load_cifar_10_data 函数首先对解压后的目录进行处理，区分出训练数据，测试数据，以及对应的标签数据。

make_data_loader 将训练数据，训练数据和标签改为训练使用的格式。并返回一个loader字典。

进行finetune训练

下面是对加载了pretrain权重的googlenet 进行filetune训练的核心逻辑

import time
import copy

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
def train_model(model, dataloaders, criterion, optimizer, num_epochs=25, is_inception=False):
    since = time.time()

    val_acc_history = []

    best_model_wts = copy.deepcopy(model.state_dict())
    best_acc = 0.0

    print("begin to train")
    for epoch in range(num_epochs):
        print('Epoch {}/{}'.format(epoch, num_epochs - 1))
        print('-' * 10)

        # Each epoch has a training and validation phase
        for phase in ['train', 'val']:
            if phase == 'train':
                model.train()  # Set model to training mode
            else:
                model.eval()   # Set model to evaluate mode

            running_loss = 0.0
            running_corrects = 0

            
            # Iterate over data.
            for inputs, labels in dataloaders[phase]:
                #print(inputs.shape)
                #print(labels.shape)
                inputs = inputs.to(device)
                labels = labels.to(device)
                

                # zero the parameter gradients
                optimizer.zero_grad()

                # forward
                # track history if only in train
                with torch.set_grad_enabled(phase == 'train'):
                    # Get model outputs and calculate loss
                    # Special case for inception because in training it has an auxiliary output. In train
                    #   mode we calculate the loss by summing the final output and the auxiliary output
                    #   but in testing we only consider the final output.
                    
                    if is_inception and phase == 'train':
                        # we don't train aux_outputs here.
                        pass
                        # From https://discuss.pytorch.org/t/how-to-optimize-inception-model-with-auxiliary-classifiers/7958
                        outputs, aux_outputs = model(inputs)
                        loss1 = criterion(outputs, labels)
                        loss2 = criterion(aux_outputs, labels)
                        loss = loss1 + 0.4*loss2
                    
                    else:
                        outputs = model(inputs)
                        #output_id = torch.max(outputs,1)[1].float()
                        #label_id = torch.max(labels,1)[1].long()
                        
                        #print(outputs)
                        #print(labels)
                        
                        loss = criterion(outputs, labels)

                    _, preds = torch.max(outputs, 1)

                    # backward + optimize only if in training phase
                    if phase == 'train':
                        loss.backward()
                        optimizer.step()

                # statistics
                running_loss += loss.item() * inputs.size(0)
                running_corrects += torch.sum(preds == labels.data)

            epoch_loss = running_loss / len(dataloaders[phase])
            epoch_acc = running_corrects.double() / len(dataloaders[phase])

            print('{} Loss: {:.4f} Acc: {:.4f}'.format(phase, epoch_loss, epoch_acc))

            # deep copy the model
            if phase == 'val' and epoch_acc > best_acc:
                best_acc = epoch_acc
                best_model_wts = copy.deepcopy(model.state_dict())
            if phase == 'val':
                val_acc_history.append(epoch_acc)

        print()

    time_elapsed = time.time() - since
    print('Training complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
    print('Best val Acc: {:4f}'.format(best_acc))

    # load best model weights
    model.load_state_dict(best_model_wts)
    return model, val_acc_history
    
criterion = nn.CrossEntropyLoss()
params_to_update = model.parameters()
print(type(params_to_update))
optimizer_ft = torch.optim.SGD(params_to_update, lr=0.001, momentum=0.9)

train_model(model, loader, criterion, optimizer_ft)

train_model函数主要逻辑也比较简单：

数据分为train 和eval两块

train阶段模型可以更新权重（进行反向传播loss.backward和权重更新optimizer.step）

eval 阶段只进行前向传播和计算准确率

整体就是pytorch训练的一般套路：

1 数据拷入device ； inputs = inputs.to(device)

2 重置梯度； optimizer.zero_grad()

3 前向传播； outputs = model(inputs)

4 计算损失；loss = criterion(outputs, labels)

5 损失反向传播； loss.backward()

6 更新权重 optimizer.step()

损失函数我们使用交叉熵损失。优化器使用SGD，动量为0.9

部分训练结果

我们这只使用了训练数据集中前100个数据进行训练。可以看到到了第18个epoch train loss下降到了1.45，val loss 却下降不明显，有可能是过拟合有可能是数据集太少。我们这里只是示意训练，后续详细调优。

<class 'generator'>
begin to train
Epoch 0/24
----------
train Loss: 2.6431 Acc: 0.0900
val Loss: 2.3888 Acc: 0.1200

Epoch 1/24
----------
train Loss: 2.6619 Acc: 0.1100
val Loss: 2.3855 Acc: 0.1300

Epoch 2/24
----------
train Loss: 2.5731 Acc: 0.1100
val Loss: 2.4109 Acc: 0.0900

Epoch 3/24
----------
train Loss: 2.5192 Acc: 0.1000
val Loss: 2.3500 Acc: 0.1200

Epoch 4/24
----------
train Loss: 2.4641 Acc: 0.0900
val Loss: 2.3360 Acc: 0.1000

.
.
.

Epoch 17/24
----------
train Loss: 1.4424 Acc: 0.4700
val Loss: 2.3126 Acc: 0.2100

Epoch 18/24
----------
train Loss: 1.4533 Acc: 0.5100
val Loss: 2.3580 Acc: 0.2300