Notebook 14: Ising Phases using Convolutional Neural Networks in PyTorch¶
Learning Goal¶
The primary goal of this notebooks is to learn how to implement a Convolutional Neural Network (CNN) using the powerful PyTorch package. It also introduces core concepts for CNNs such as convolutional and pooling layers and padding.
Overview¶
In this notebook, we will write a simple convolutional neural network (CNN) in Pytorch for classifying phases for the Ising Model. We will consider perhaps the simplest CNN: a single convolutional layer with depth $N \in \{1,5,10,20,50\}$. This will introduce the power of the Pytorch framework to make dynamic graphs.
We will use this notebook to characterize samples drawn from the 2D Ising model at various temperatures. This is the same dataset that has been used in all earlier examples. Recall that the critical temperature for the Ising model is $T_c=2.26$.
In [1]:
from __future__ import print_function, division
import os,sys
import numpy as np
import torch # pytorch package, allows using GPUs
# fix seed
seed=17
np.random.seed(seed)
torch.manual_seed(seed)
Out[1]:
<torch._C.Generator at 0x10fee8bd0>
Structure of the Procedure¶
Constructing a Deep Neural Network to solve ML problems is a multiple-stage process. Quite generally, one can identify the key steps as follows:
- step 1: Load and process the data
- step 2: Define the model and its architecture
- step 3: Choose the optimizer and the cost function
- step 4: Train the model
- step 5: Evaluate the model performance on the unseen test data
- step 6: Modify the hyperparameters to optimize performance for the specific data set
Below, we sometimes combine some of these steps together for convenience.
Step 1: Load and Process the Ising Dataset¶
We start by defining the dataset class for Pytorch.
We have three types of samples in the Ising dataset: samples drawn from deep in the disordered phase, samples drawn from the ordered phase, and samples drawn from near the critical phase which we do not use for training. The goal is to classify whether a sample comes from $T>T_c$ or $T<T_c$.
There is standard way to load data when using the PyTorch package, which we discussed in the DNN example for the SUSY dataset. Here, we just switch to the 2D-Ising data instead.
To proceed, download the Ising dataset and insert the proper path_to_data in the script.
In [2]:
from torchvision import datasets # load data
class Ising_Dataset(torch.utils.data.Dataset):
"""Ising pytorch dataset."""
def __init__(self, data_type, transform=False):
"""
Args:
data_type (string): `train`, `test` or `critical`: creates data_loader
transform (callable, optional): Optional transform to be applied on a sample.
"""
from sklearn.model_selection import train_test_split
import collections
import pickle as pickle
L=40 # linear system size
T=np.linspace(0.25,4.0,16) # temperatures
T_c=2.26 # critical temperature in the TD limit
# path to data directory
path_to_data=os.path.expanduser('~')+'/Dropbox/MachineLearningReview/Datasets/isingMC/'
# load data
file_name = "Ising2DFM_reSample_L40_T=All.pkl" # this file contains 16*10000 samples taken in T=np.arange(0.25,4.0001,0.25)
data = pickle.load(open(path_to_data+file_name,'rb')) # pickle reads the file and returns the Python object (1D array, compressed bits)
data = np.unpackbits(data).reshape(-1, 1600) # Decompress array and reshape for convenience
data=data.astype('int')
data[np.where(data==0)]=-1 # map 0 state to -1 (Ising variable can take values +/-1)
file_name = "Ising2DFM_reSample_L40_T=All_labels.pkl" # this file contains 16*10000 samples taken in T=np.arange(0.25,4.0001,0.25)
labels = pickle.load(open(path_to_data+file_name,'rb')) # pickle reads the file and returns the Python object (here just a 1D array with the binary labels)
# divide data into ordered, critical and disordered
X_ordered=data[:70000,:]
Y_ordered=labels[:70000]
X_critical=data[70000:100000,:]
Y_critical=labels[70000:100000]
X_disordered=data[100000:,:]
Y_disordered=labels[100000:]
del data,labels
# define training, critical and test data sets
X=np.concatenate((X_ordered,X_disordered)) #np.concatenate((X_ordered,X_critical,X_disordered))
Y=np.concatenate((Y_ordered,Y_disordered)) #np.concatenate((Y_ordered,Y_critical,Y_disordered))
# pick random data points from ordered and disordered states to create the training and test sets
X_train,X_test,Y_train,Y_test=train_test_split(X,Y,test_size=0.2,train_size=0.8)
if data_type=='train':
X=X_train
Y=Y_train
print("Training on 80 percent of examples")
if data_type=='test':
X=X_test
Y=Y_test
print("Testing on 20 percent of examples")
if data_type=='critical':
X=X_critical
Y=Y_critical
print("Predicting on %i critical examples"%len(Y_critical))
# reshape data back to original 2D-array form
X=X.reshape(X.shape[0],40,40)
# these are necessary attributes in dataset class and must be assigned
self.data=(X,Y)
self.transform = transform
# override __len__ and __getitem__ of the Dataset() class
def __len__(self):
return len(self.data[1])
def __getitem__(self, idx):
sample=(self.data[0][idx,...],self.data[1][idx])
if self.transform:
sample=self.transform(sample)
return sample
def load_data(kwargs):
# kwargs: CUDA arguments, if enabled
# load and noralise train,test, and data
train_loader = torch.utils.data.DataLoader(
Ising_Dataset(data_type='train'),
batch_size=args.batch_size, shuffle=True, **kwargs)
test_loader = torch.utils.data.DataLoader(
Ising_Dataset(data_type='test'),
batch_size=args.test_batch_size, shuffle=True, **kwargs)
critical_loader = torch.utils.data.DataLoader(
Ising_Dataset(data_type='critical'),
batch_size=args.test_batch_size, shuffle=True, **kwargs)
return train_loader, test_loader, critical_loader
Step 2: Define the Neural Net and its Architecture¶
Similar to the discussion in the SUSY DNN notebook, we then define the architecture of the neural net in the model class which contains the forward function method that tells us how to produce the output given some input. The backpropagaiton algorithm is implemented automatically by the Pytorch package.
Recall that a CNN is composed of convolutional layers, max-pool layer, often followed by a fully connected layer and then the classifier. In the architecture below, we start with a convolutional layer that takes as an input a layer with $D_{in}=1$ with height and width $H=W=40$, a receptive field or filter size of $2 \times 2$, and depth $N$ (there are $N$ layers). We also add a padding of zeros on both sides of the image. This convolutional layer can be summarized by the four numbers $[N,D_{in},H,W]=[N,1,41,41]$. This is then fed into a $2 \times 2$ maxpool layer which results in layer of size $[N,1,20,20]$. This layer is then hooked up to a linear layer that takes as an [input,output] of the form [N*20*20*1,2] since there are $2$ classes (corresponding to the ordered and disordered phases). We use a logistic (softmax) classifier as the output and train using various optimizers.
In [3]:
import torch.nn as nn # construct NN
class model(nn.Module):
# create convolutional net
def __init__(self, N=10, L=40):
# inherit attributes and methods of nn.Module
super(model, self).__init__()
# create convolutional layer with input depth 1 and output depth N
self.conv1 = nn.Conv2d(1, N, kernel_size=2, padding=1)
# batch norm layer takes Depth
self.bn1=nn.BatchNorm2d(N)
# create fully connected layer after maxpool operation reduced 40->18
self.fc1 = nn.Linear(20*20*N, 2)
self.N=N
self.L=L
print("The number of neurons in CNN layer is %i"%(N))
def forward(self, x):
#Unsqueeze command indicates one channel and turns x.shape from (:,40,40) to (:,1, 40,40)
x=F.relu(self.conv1(torch.unsqueeze(x,1).float()))
#print(x.shape) often useful to look at shapes for debugging
x = F.max_pool2d(x,2)
#print(x.shape)
x=self.bn1(x) # largely unnecessary and here just for pedagogical purposes
return F.log_softmax(self.fc1(x.view(-1,20*20*self.N)), dim=1)
Define the train and test functions¶
These are very standard functions for going over data to train and evaluate the model.
Since we will be testing the CNN performance on both the test and the critical data, the test function accepts two arguments: data_loader and verbose to allow control over the input data and the printing messages.
In [4]:
def train(epoch):
# these are very standard functions for going over data to train
CNN.train() # effects Dropout and BatchNorm layers
for batch_idx, (data, target) in enumerate(train_loader):
if args.cuda:
data, target = data.cuda(), target.cuda()
optimizer.zero_grad()
output = CNN(data)
loss = F.nll_loss(output, target)
loss.backward()
optimizer.step()
if batch_idx % args.log_interval == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
epoch, batch_idx * len(data), len(train_loader.dataset),
100. * batch_idx / len(train_loader), loss.item()))
def test(data_loader,verbose='Test'):
# these are very standard functions for evaluating data
CNN.eval() # effects Dropout and BatchNorm layers
test_loss = 0
correct = 0
for data, target in data_loader:
if args.cuda:
data, target = data.cuda(), target.cuda()
output = CNN(data)
test_loss += F.nll_loss(output, target, size_average=False).item() # sum up batch loss
pred = output.data.max(1, keepdim=True)[1] # get the index of the max log-probability
correct += pred.eq(target.data.view_as(pred)).cpu().sum().item()
test_loss /= len(data_loader.dataset)
print('\n'+verbose+' set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'.format(
test_loss, correct, len(data_loader.dataset),
100. * correct / len(data_loader.dataset)))
accuracy=100. * correct / len(data_loader.dataset)
return(accuracy)
Define Model Parameters¶
Next we define the training settings. This proceeds in the same way as in Notebook 13 on the SUSY dataset, except we now also show how to turn on the cuda library option of PyTorch which enables parallel coputations (whenever resources for this are available).
In [5]:
import argparse # handles arguments
import sys; sys.argv=['']; del sys # required to use parser in jupyter notebooks
# training settings
parser = argparse.ArgumentParser(description='PyTorch Convmodel Ising Example')
parser.add_argument('--batch-size', type=int, default=64, metavar='N',
help='input batch size for training (default: 64)')
parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
help='input batch size for testing (default: 1000)')
parser.add_argument('--epochs', type=int, default=10, metavar='N',
help='number of epochs to train (default: 10)')
parser.add_argument('--lr', type=float, default=0.01, metavar='LR',
help='learning rate (default: 0.01)')
parser.add_argument('--momentum', type=float, default=0.5, metavar='M',
help='SGD momentum (default: 0.5)')
parser.add_argument('--no-cuda', action='store_true', default=False,
help='disables CUDA training')
parser.add_argument('--seed', type=int, default=1, metavar='S',
help='random seed (default: 1)')
parser.add_argument('--log-interval', type=int, default=10, metavar='N',
help='how many batches to wait before logging training status')
args = parser.parse_args()
args.epochs=5
args.cuda = not args.no_cuda and torch.cuda.is_available()
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
cuda_kwargs = {'num_workers': 1, 'pin_memory': True} if args.cuda else {}
Steps 3+4+5: Choose the Optimizer and the Cost Function. Train and Evaluate the Model.¶
In [6]:
import torch.nn.functional as F # implements forward and backward definitions of an autograd operation
import torch.optim as optim # different update rules such as SGD, Nesterov-SGD, Adam, RMSProp, etc
# load data
train_loader, test_loader, critical_loader=load_data(cuda_kwargs)
test_array=[]
critical_array=[]
# create array of depth of convolutional layer
N_array=[1,5,10,20,50]
# loop over depths
for N in N_array:
CNN = model(N=N)
if args.cuda:
CNN.cuda()
# negative log-likelihood (nll) loss for training: takes class labels NOT one-hot vectors!
criterion = F.nll_loss
# define SGD optimizer
optimizer = optim.SGD(CNN.parameters(), lr=args.lr, momentum=args.momentum)
#optimizer = optim.Adam(DNN.parameters(), lr=0.001, betas=(0.9, 0.999))
# train the CNN and test its performance at each epoch
for epoch in range(1, args.epochs + 1):
train(epoch)
if epoch==args.epochs:
test_array.append(test(test_loader,verbose='Test'))
critical_array.append(test(critical_loader,verbose='Critical'))
else:
test(test_loader,verbose='Test')
test(critical_loader,verbose='Critical')
print(test_array)
print(critical_array)
Training on 80 percent of examples Testing on 20 percent of examples Predicting on 30000 critical examples The number of neurons in CNN layer is 1 Train Epoch: 1 [0/104000 (0%)] Loss: 0.662742 Train Epoch: 1 [640/104000 (1%)] Loss: 0.109889 Train Epoch: 1 [1280/104000 (1%)] Loss: 0.045462 Train Epoch: 1 [1920/104000 (2%)] Loss: 0.050722 Train Epoch: 1 [2560/104000 (2%)] Loss: 0.020955 Train Epoch: 1 [3200/104000 (3%)] Loss: 0.016692 Train Epoch: 1 [3840/104000 (4%)] Loss: 0.019543 Train Epoch: 1 [4480/104000 (4%)] Loss: 0.012761 Train Epoch: 1 [5120/104000 (5%)] Loss: 0.017204 Train Epoch: 1 [5760/104000 (6%)] Loss: 0.011533 Train Epoch: 1 [6400/104000 (6%)] Loss: 0.007559 Train Epoch: 1 [7040/104000 (7%)] Loss: 0.003841 Train Epoch: 1 [7680/104000 (7%)] Loss: 0.004554 Train Epoch: 1 [8320/104000 (8%)] Loss: 0.008697 Train Epoch: 1 [8960/104000 (9%)] Loss: 0.005727 Train Epoch: 1 [9600/104000 (9%)] Loss: 0.016714 Train Epoch: 1 [10240/104000 (10%)] Loss: 0.002938 Train Epoch: 1 [10880/104000 (10%)] Loss: 0.009886 Train Epoch: 1 [11520/104000 (11%)] Loss: 0.006929 Train Epoch: 1 [12160/104000 (12%)] Loss: 0.002550 Train Epoch: 1 [12800/104000 (12%)] Loss: 0.008572 Train Epoch: 1 [13440/104000 (13%)] Loss: 0.002893 Train Epoch: 1 [14080/104000 (14%)] Loss: 0.001733 Train Epoch: 1 [14720/104000 (14%)] Loss: 0.009416 Train Epoch: 1 [15360/104000 (15%)] Loss: 0.002594 Train Epoch: 1 [16000/104000 (15%)] Loss: 0.006668 Train Epoch: 1 [16640/104000 (16%)] Loss: 0.001514 Train Epoch: 1 [17280/104000 (17%)] Loss: 0.005911 Train Epoch: 1 [17920/104000 (17%)] Loss: 0.004004 Train Epoch: 1 [18560/104000 (18%)] Loss: 0.004119 Train Epoch: 1 [19200/104000 (18%)] Loss: 0.015166 Train Epoch: 1 [19840/104000 (19%)] Loss: 0.003167 Train Epoch: 1 [20480/104000 (20%)] Loss: 0.002569 Train Epoch: 1 [21120/104000 (20%)] Loss: 0.003589 Train Epoch: 1 [21760/104000 (21%)] Loss: 0.002717 Train Epoch: 1 [22400/104000 (22%)] Loss: 0.004898 Train Epoch: 1 [23040/104000 (22%)] Loss: 0.002880 Train Epoch: 1 [23680/104000 (23%)] Loss: 0.004688 Train Epoch: 1 [24320/104000 (23%)] Loss: 0.000671 Train Epoch: 1 [24960/104000 (24%)] Loss: 0.004286 Train Epoch: 1 [25600/104000 (25%)] Loss: 0.001073 Train Epoch: 1 [26240/104000 (25%)] Loss: 0.002328 Train Epoch: 1 [26880/104000 (26%)] Loss: 0.002988 Train Epoch: 1 [27520/104000 (26%)] Loss: 0.001294 Train Epoch: 1 [28160/104000 (27%)] Loss: 0.052510 Train Epoch: 1 [28800/104000 (28%)] Loss: 0.000678 Train Epoch: 1 [29440/104000 (28%)] Loss: 0.000652 Train Epoch: 1 [30080/104000 (29%)] Loss: 0.003534 Train Epoch: 1 [30720/104000 (30%)] Loss: 0.012102 Train Epoch: 1 [31360/104000 (30%)] Loss: 0.000991 Train Epoch: 1 [32000/104000 (31%)] Loss: 0.002256 Train Epoch: 1 [32640/104000 (31%)] Loss: 0.004379 Train Epoch: 1 [33280/104000 (32%)] Loss: 0.001981 Train Epoch: 1 [33920/104000 (33%)] Loss: 0.001887 Train Epoch: 1 [34560/104000 (33%)] Loss: 0.004612 Train Epoch: 1 [35200/104000 (34%)] Loss: 0.005662 Train Epoch: 1 [35840/104000 (34%)] Loss: 0.011141 Train Epoch: 1 [36480/104000 (35%)] Loss: 0.001433 Train Epoch: 1 [37120/104000 (36%)] Loss: 0.002003 Train Epoch: 1 [37760/104000 (36%)] Loss: 0.000672 Train Epoch: 1 [38400/104000 (37%)] Loss: 0.001759 Train Epoch: 1 [39040/104000 (38%)] Loss: 0.002058 Train Epoch: 1 [39680/104000 (38%)] Loss: 0.001550 Train Epoch: 1 [40320/104000 (39%)] Loss: 0.001455 Train Epoch: 1 [40960/104000 (39%)] Loss: 0.000544 Train Epoch: 1 [41600/104000 (40%)] Loss: 0.004700 Train Epoch: 1 [42240/104000 (41%)] Loss: 0.000772 Train Epoch: 1 [42880/104000 (41%)] Loss: 0.001735 Train Epoch: 1 [43520/104000 (42%)] Loss: 0.001120 Train Epoch: 1 [44160/104000 (42%)] Loss: 0.001542 Train Epoch: 1 [44800/104000 (43%)] Loss: 0.009389 Train Epoch: 1 [45440/104000 (44%)] Loss: 0.000525 Train Epoch: 1 [46080/104000 (44%)] Loss: 0.000306 Train Epoch: 1 [46720/104000 (45%)] Loss: 0.007553 Train Epoch: 1 [47360/104000 (46%)] Loss: 0.000785 Train Epoch: 1 [48000/104000 (46%)] Loss: 0.003900 Train Epoch: 1 [48640/104000 (47%)] Loss: 0.001991 Train Epoch: 1 [49280/104000 (47%)] Loss: 0.002793 Train Epoch: 1 [49920/104000 (48%)] Loss: 0.020791 Train Epoch: 1 [50560/104000 (49%)] Loss: 0.000786 Train Epoch: 1 [51200/104000 (49%)] Loss: 0.000717 Train Epoch: 1 [51840/104000 (50%)] Loss: 0.000714 Train Epoch: 1 [52480/104000 (50%)] Loss: 0.001216 Train Epoch: 1 [53120/104000 (51%)] Loss: 0.000183 Train Epoch: 1 [53760/104000 (52%)] Loss: 0.000735 Train Epoch: 1 [54400/104000 (52%)] Loss: 0.002544 Train Epoch: 1 [55040/104000 (53%)] Loss: 0.000421 Train Epoch: 1 [55680/104000 (54%)] Loss: 0.000800 Train Epoch: 1 [56320/104000 (54%)] Loss: 0.008846 Train Epoch: 1 [56960/104000 (55%)] Loss: 0.000468 Train Epoch: 1 [57600/104000 (55%)] Loss: 0.001781 Train Epoch: 1 [58240/104000 (56%)] Loss: 0.000885 Train Epoch: 1 [58880/104000 (57%)] Loss: 0.001745 Train Epoch: 1 [59520/104000 (57%)] Loss: 0.000826 Train Epoch: 1 [60160/104000 (58%)] Loss: 0.000531 Train Epoch: 1 [60800/104000 (58%)] Loss: 0.003100 Train Epoch: 1 [61440/104000 (59%)] Loss: 0.000649 Train Epoch: 1 [62080/104000 (60%)] Loss: 0.000953 Train Epoch: 1 [62720/104000 (60%)] Loss: 0.000340 Train Epoch: 1 [63360/104000 (61%)] Loss: 0.003273 Train Epoch: 1 [64000/104000 (62%)] Loss: 0.001740 Train Epoch: 1 [64640/104000 (62%)] Loss: 0.000066 Train Epoch: 1 [65280/104000 (63%)] Loss: 0.001644 Train Epoch: 1 [65920/104000 (63%)] Loss: 0.001525 Train Epoch: 1 [66560/104000 (64%)] Loss: 0.000233 Train Epoch: 1 [67200/104000 (65%)] Loss: 0.001457 Train Epoch: 1 [67840/104000 (65%)] Loss: 0.000194 Train Epoch: 1 [68480/104000 (66%)] Loss: 0.001445 Train Epoch: 1 [69120/104000 (66%)] Loss: 0.002877 Train Epoch: 1 [69760/104000 (67%)] Loss: 0.000257 Train Epoch: 1 [70400/104000 (68%)] Loss: 0.000617 Train Epoch: 1 [71040/104000 (68%)] Loss: 0.000587 Train Epoch: 1 [71680/104000 (69%)] Loss: 0.000208 Train Epoch: 1 [72320/104000 (70%)] Loss: 0.001745 Train Epoch: 1 [72960/104000 (70%)] Loss: 0.000251 Train Epoch: 1 [73600/104000 (71%)] Loss: 0.000186 Train Epoch: 1 [74240/104000 (71%)] Loss: 0.000555 Train Epoch: 1 [74880/104000 (72%)] Loss: 0.000174 Train Epoch: 1 [75520/104000 (73%)] Loss: 0.000444 Train Epoch: 1 [76160/104000 (73%)] Loss: 0.000899 Train Epoch: 1 [76800/104000 (74%)] Loss: 0.000435 Train Epoch: 1 [77440/104000 (74%)] Loss: 0.000337 Train Epoch: 1 [78080/104000 (75%)] Loss: 0.001668 Train Epoch: 1 [78720/104000 (76%)] Loss: 0.000330 Train Epoch: 1 [79360/104000 (76%)] Loss: 0.000514 Train Epoch: 1 [80000/104000 (77%)] Loss: 0.006597 Train Epoch: 1 [80640/104000 (78%)] Loss: 0.001635 Train Epoch: 1 [81280/104000 (78%)] Loss: 0.000408 Train Epoch: 1 [81920/104000 (79%)] Loss: 0.000226 Train Epoch: 1 [82560/104000 (79%)] Loss: 0.001308 Train Epoch: 1 [83200/104000 (80%)] Loss: 0.000930 Train Epoch: 1 [83840/104000 (81%)] Loss: 0.002042 Train Epoch: 1 [84480/104000 (81%)] Loss: 0.000472 Train Epoch: 1 [85120/104000 (82%)] Loss: 0.000097 Train Epoch: 1 [85760/104000 (82%)] Loss: 0.000263 Train Epoch: 1 [86400/104000 (83%)] Loss: 0.000370 Train Epoch: 1 [87040/104000 (84%)] Loss: 0.000702 Train Epoch: 1 [87680/104000 (84%)] Loss: 0.000267 Train Epoch: 1 [88320/104000 (85%)] Loss: 0.000295 Train Epoch: 1 [88960/104000 (86%)] Loss: 0.000098 Train Epoch: 1 [89600/104000 (86%)] Loss: 0.000813 Train Epoch: 1 [90240/104000 (87%)] Loss: 0.001157 Train Epoch: 1 [90880/104000 (87%)] Loss: 0.000578 Train Epoch: 1 [91520/104000 (88%)] Loss: 0.005544 Train Epoch: 1 [92160/104000 (89%)] Loss: 0.000640 Train Epoch: 1 [92800/104000 (89%)] Loss: 0.000296 Train Epoch: 1 [93440/104000 (90%)] Loss: 0.000367 Train Epoch: 1 [94080/104000 (90%)] Loss: 0.000388 Train Epoch: 1 [94720/104000 (91%)] Loss: 0.002296 Train Epoch: 1 [95360/104000 (92%)] Loss: 0.002912 Train Epoch: 1 [96000/104000 (92%)] Loss: 0.000297 Train Epoch: 1 [96640/104000 (93%)] Loss: 0.000208 Train Epoch: 1 [97280/104000 (94%)] Loss: 0.001088 Train Epoch: 1 [97920/104000 (94%)] Loss: 0.000779 Train Epoch: 1 [98560/104000 (95%)] Loss: 0.000636 Train Epoch: 1 [99200/104000 (95%)] Loss: 0.000509 Train Epoch: 1 [99840/104000 (96%)] Loss: 0.001388 Train Epoch: 1 [100480/104000 (97%)] Loss: 0.002664 Train Epoch: 1 [101120/104000 (97%)] Loss: 0.000189 Train Epoch: 1 [101760/104000 (98%)] Loss: 0.000681 Train Epoch: 1 [102400/104000 (98%)] Loss: 0.000616 Train Epoch: 1 [103040/104000 (99%)] Loss: 0.000326 Train Epoch: 1 [103680/104000 (100%)] Loss: 0.000554
/Users/mgbukov/miniconda3/envs/mlreview/lib/python3.6/site-packages/torch/nn/functional.py:52: UserWarning: size_average and reduce args will be deprecated, please use reduction='sum' instead. warnings.warn(warning.format(ret))
Test set: Average loss: 0.0005, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.8165, Accuracy: 23875/30000 (80%) Train Epoch: 2 [0/104000 (0%)] Loss: 0.000207 Train Epoch: 2 [640/104000 (1%)] Loss: 0.003293 Train Epoch: 2 [1280/104000 (1%)] Loss: 0.002911 Train Epoch: 2 [1920/104000 (2%)] Loss: 0.000456 Train Epoch: 2 [2560/104000 (2%)] Loss: 0.000537 Train Epoch: 2 [3200/104000 (3%)] Loss: 0.000169 Train Epoch: 2 [3840/104000 (4%)] Loss: 0.004852 Train Epoch: 2 [4480/104000 (4%)] Loss: 0.000377 Train Epoch: 2 [5120/104000 (5%)] Loss: 0.000677 Train Epoch: 2 [5760/104000 (6%)] Loss: 0.000874 Train Epoch: 2 [6400/104000 (6%)] Loss: 0.000330 Train Epoch: 2 [7040/104000 (7%)] Loss: 0.000055 Train Epoch: 2 [7680/104000 (7%)] Loss: 0.000904 Train Epoch: 2 [8320/104000 (8%)] Loss: 0.000891 Train Epoch: 2 [8960/104000 (9%)] Loss: 0.001143 Train Epoch: 2 [9600/104000 (9%)] Loss: 0.000845 Train Epoch: 2 [10240/104000 (10%)] Loss: 0.000624 Train Epoch: 2 [10880/104000 (10%)] Loss: 0.000672 Train Epoch: 2 [11520/104000 (11%)] Loss: 0.000998 Train Epoch: 2 [12160/104000 (12%)] Loss: 0.000145 Train Epoch: 2 [12800/104000 (12%)] Loss: 0.000117 Train Epoch: 2 [13440/104000 (13%)] Loss: 0.000100 Train Epoch: 2 [14080/104000 (14%)] Loss: 0.000311 Train Epoch: 2 [14720/104000 (14%)] Loss: 0.001324 Train Epoch: 2 [15360/104000 (15%)] Loss: 0.000712 Train Epoch: 2 [16000/104000 (15%)] Loss: 0.000173 Train Epoch: 2 [16640/104000 (16%)] Loss: 0.000301 Train Epoch: 2 [17280/104000 (17%)] Loss: 0.000548 Train Epoch: 2 [17920/104000 (17%)] Loss: 0.000567 Train Epoch: 2 [18560/104000 (18%)] Loss: 0.000097 Train Epoch: 2 [19200/104000 (18%)] Loss: 0.001975 Train Epoch: 2 [19840/104000 (19%)] Loss: 0.000741 Train Epoch: 2 [20480/104000 (20%)] Loss: 0.000132 Train Epoch: 2 [21120/104000 (20%)] Loss: 0.000875 Train Epoch: 2 [21760/104000 (21%)] Loss: 0.003284 Train Epoch: 2 [22400/104000 (22%)] Loss: 0.000269 Train Epoch: 2 [23040/104000 (22%)] Loss: 0.000087 Train Epoch: 2 [23680/104000 (23%)] Loss: 0.000139 Train Epoch: 2 [24320/104000 (23%)] Loss: 0.001205 Train Epoch: 2 [24960/104000 (24%)] Loss: 0.000231 Train Epoch: 2 [25600/104000 (25%)] Loss: 0.000141 Train Epoch: 2 [26240/104000 (25%)] Loss: 0.000569 Train Epoch: 2 [26880/104000 (26%)] Loss: 0.000188 Train Epoch: 2 [27520/104000 (26%)] Loss: 0.000104 Train Epoch: 2 [28160/104000 (27%)] Loss: 0.000142 Train Epoch: 2 [28800/104000 (28%)] Loss: 0.000027 Train Epoch: 2 [29440/104000 (28%)] Loss: 0.000069 Train Epoch: 2 [30080/104000 (29%)] Loss: 0.001534 Train Epoch: 2 [30720/104000 (30%)] Loss: 0.000342 Train Epoch: 2 [31360/104000 (30%)] Loss: 0.000902 Train Epoch: 2 [32000/104000 (31%)] Loss: 0.000623 Train Epoch: 2 [32640/104000 (31%)] Loss: 0.000091 Train Epoch: 2 [33280/104000 (32%)] Loss: 0.000142 Train Epoch: 2 [33920/104000 (33%)] Loss: 0.003602 Train Epoch: 2 [34560/104000 (33%)] Loss: 0.000527 Train Epoch: 2 [35200/104000 (34%)] Loss: 0.000956 Train Epoch: 2 [35840/104000 (34%)] Loss: 0.000588 Train Epoch: 2 [36480/104000 (35%)] Loss: 0.001144 Train Epoch: 2 [37120/104000 (36%)] Loss: 0.000418 Train Epoch: 2 [37760/104000 (36%)] Loss: 0.002153 Train Epoch: 2 [38400/104000 (37%)] Loss: 0.000243 Train Epoch: 2 [39040/104000 (38%)] Loss: 0.001616 Train Epoch: 2 [39680/104000 (38%)] Loss: 0.000081 Train Epoch: 2 [40320/104000 (39%)] Loss: 0.000097 Train Epoch: 2 [40960/104000 (39%)] Loss: 0.000208 Train Epoch: 2 [41600/104000 (40%)] Loss: 0.000639 Train Epoch: 2 [42240/104000 (41%)] Loss: 0.001528 Train Epoch: 2 [42880/104000 (41%)] Loss: 0.000136 Train Epoch: 2 [43520/104000 (42%)] Loss: 0.000272 Train Epoch: 2 [44160/104000 (42%)] Loss: 0.000405 Train Epoch: 2 [44800/104000 (43%)] Loss: 0.000983 Train Epoch: 2 [45440/104000 (44%)] Loss: 0.000348 Train Epoch: 2 [46080/104000 (44%)] Loss: 0.000270 Train Epoch: 2 [46720/104000 (45%)] Loss: 0.001273 Train Epoch: 2 [47360/104000 (46%)] Loss: 0.000122 Train Epoch: 2 [48000/104000 (46%)] Loss: 0.000065 Train Epoch: 2 [48640/104000 (47%)] Loss: 0.000083 Train Epoch: 2 [49280/104000 (47%)] Loss: 0.000129 Train Epoch: 2 [49920/104000 (48%)] Loss: 0.000709 Train Epoch: 2 [50560/104000 (49%)] Loss: 0.000526 Train Epoch: 2 [51200/104000 (49%)] Loss: 0.000125 Train Epoch: 2 [51840/104000 (50%)] Loss: 0.000151 Train Epoch: 2 [52480/104000 (50%)] Loss: 0.000048 Train Epoch: 2 [53120/104000 (51%)] Loss: 0.000072 Train Epoch: 2 [53760/104000 (52%)] Loss: 0.000176 Train Epoch: 2 [54400/104000 (52%)] Loss: 0.001160 Train Epoch: 2 [55040/104000 (53%)] Loss: 0.003954 Train Epoch: 2 [55680/104000 (54%)] Loss: 0.004039 Train Epoch: 2 [56320/104000 (54%)] Loss: 0.000845 Train Epoch: 2 [56960/104000 (55%)] Loss: 0.000327 Train Epoch: 2 [57600/104000 (55%)] Loss: 0.000085 Train Epoch: 2 [58240/104000 (56%)] Loss: 0.000056 Train Epoch: 2 [58880/104000 (57%)] Loss: 0.000969 Train Epoch: 2 [59520/104000 (57%)] Loss: 0.000158 Train Epoch: 2 [60160/104000 (58%)] Loss: 0.000294 Train Epoch: 2 [60800/104000 (58%)] Loss: 0.000255 Train Epoch: 2 [61440/104000 (59%)] Loss: 0.000088 Train Epoch: 2 [62080/104000 (60%)] Loss: 0.003106 Train Epoch: 2 [62720/104000 (60%)] Loss: 0.004540 Train Epoch: 2 [63360/104000 (61%)] Loss: 0.000265 Train Epoch: 2 [64000/104000 (62%)] Loss: 0.000957 Train Epoch: 2 [64640/104000 (62%)] Loss: 0.000135 Train Epoch: 2 [65280/104000 (63%)] Loss: 0.000137 Train Epoch: 2 [65920/104000 (63%)] Loss: 0.000109 Train Epoch: 2 [66560/104000 (64%)] Loss: 0.000603 Train Epoch: 2 [67200/104000 (65%)] Loss: 0.000102 Train Epoch: 2 [67840/104000 (65%)] Loss: 0.000302 Train Epoch: 2 [68480/104000 (66%)] Loss: 0.000063 Train Epoch: 2 [69120/104000 (66%)] Loss: 0.000650 Train Epoch: 2 [69760/104000 (67%)] Loss: 0.000157 Train Epoch: 2 [70400/104000 (68%)] Loss: 0.000306 Train Epoch: 2 [71040/104000 (68%)] Loss: 0.033093 Train Epoch: 2 [71680/104000 (69%)] Loss: 0.000443 Train Epoch: 2 [72320/104000 (70%)] Loss: 0.000308 Train Epoch: 2 [72960/104000 (70%)] Loss: 0.000392 Train Epoch: 2 [73600/104000 (71%)] Loss: 0.000295 Train Epoch: 2 [74240/104000 (71%)] Loss: 0.001193 Train Epoch: 2 [74880/104000 (72%)] Loss: 0.003020 Train Epoch: 2 [75520/104000 (73%)] Loss: 0.000096 Train Epoch: 2 [76160/104000 (73%)] Loss: 0.000202 Train Epoch: 2 [76800/104000 (74%)] Loss: 0.000464 Train Epoch: 2 [77440/104000 (74%)] Loss: 0.000096 Train Epoch: 2 [78080/104000 (75%)] Loss: 0.000068 Train Epoch: 2 [78720/104000 (76%)] Loss: 0.002312 Train Epoch: 2 [79360/104000 (76%)] Loss: 0.000055 Train Epoch: 2 [80000/104000 (77%)] Loss: 0.000171 Train Epoch: 2 [80640/104000 (78%)] Loss: 0.002382 Train Epoch: 2 [81280/104000 (78%)] Loss: 0.000563 Train Epoch: 2 [81920/104000 (79%)] Loss: 0.000178 Train Epoch: 2 [82560/104000 (79%)] Loss: 0.000115 Train Epoch: 2 [83200/104000 (80%)] Loss: 0.000178 Train Epoch: 2 [83840/104000 (81%)] Loss: 0.000097 Train Epoch: 2 [84480/104000 (81%)] Loss: 0.000127 Train Epoch: 2 [85120/104000 (82%)] Loss: 0.001672 Train Epoch: 2 [85760/104000 (82%)] Loss: 0.000080 Train Epoch: 2 [86400/104000 (83%)] Loss: 0.000045 Train Epoch: 2 [87040/104000 (84%)] Loss: 0.000213 Train Epoch: 2 [87680/104000 (84%)] Loss: 0.000210 Train Epoch: 2 [88320/104000 (85%)] Loss: 0.000272 Train Epoch: 2 [88960/104000 (86%)] Loss: 0.001581 Train Epoch: 2 [89600/104000 (86%)] Loss: 0.000713 Train Epoch: 2 [90240/104000 (87%)] Loss: 0.000070 Train Epoch: 2 [90880/104000 (87%)] Loss: 0.002180 Train Epoch: 2 [91520/104000 (88%)] Loss: 0.000037 Train Epoch: 2 [92160/104000 (89%)] Loss: 0.000250 Train Epoch: 2 [92800/104000 (89%)] Loss: 0.000103 Train Epoch: 2 [93440/104000 (90%)] Loss: 0.000222 Train Epoch: 2 [94080/104000 (90%)] Loss: 0.000026 Train Epoch: 2 [94720/104000 (91%)] Loss: 0.000986 Train Epoch: 2 [95360/104000 (92%)] Loss: 0.000350 Train Epoch: 2 [96000/104000 (92%)] Loss: 0.000084 Train Epoch: 2 [96640/104000 (93%)] Loss: 0.000553 Train Epoch: 2 [97280/104000 (94%)] Loss: 0.000681 Train Epoch: 2 [97920/104000 (94%)] Loss: 0.000453 Train Epoch: 2 [98560/104000 (95%)] Loss: 0.000199 Train Epoch: 2 [99200/104000 (95%)] Loss: 0.000310 Train Epoch: 2 [99840/104000 (96%)] Loss: 0.000116 Train Epoch: 2 [100480/104000 (97%)] Loss: 0.000248 Train Epoch: 2 [101120/104000 (97%)] Loss: 0.000102 Train Epoch: 2 [101760/104000 (98%)] Loss: 0.000298 Train Epoch: 2 [102400/104000 (98%)] Loss: 0.000583 Train Epoch: 2 [103040/104000 (99%)] Loss: 0.000026 Train Epoch: 2 [103680/104000 (100%)] Loss: 0.000239 Test set: Average loss: 0.0003, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.9904, Accuracy: 23634/30000 (79%) Train Epoch: 3 [0/104000 (0%)] Loss: 0.000187 Train Epoch: 3 [640/104000 (1%)] Loss: 0.000054 Train Epoch: 3 [1280/104000 (1%)] Loss: 0.000154 Train Epoch: 3 [1920/104000 (2%)] Loss: 0.000851 Train Epoch: 3 [2560/104000 (2%)] Loss: 0.000049 Train Epoch: 3 [3200/104000 (3%)] Loss: 0.000274 Train Epoch: 3 [3840/104000 (4%)] Loss: 0.000528 Train Epoch: 3 [4480/104000 (4%)] Loss: 0.000440 Train Epoch: 3 [5120/104000 (5%)] Loss: 0.001590 Train Epoch: 3 [5760/104000 (6%)] Loss: 0.003739 Train Epoch: 3 [6400/104000 (6%)] Loss: 0.000099 Train Epoch: 3 [7040/104000 (7%)] Loss: 0.000564 Train Epoch: 3 [7680/104000 (7%)] Loss: 0.000268 Train Epoch: 3 [8320/104000 (8%)] Loss: 0.000250 Train Epoch: 3 [8960/104000 (9%)] Loss: 0.000432 Train Epoch: 3 [9600/104000 (9%)] Loss: 0.000188 Train Epoch: 3 [10240/104000 (10%)] Loss: 0.000163 Train Epoch: 3 [10880/104000 (10%)] Loss: 0.000407 Train Epoch: 3 [11520/104000 (11%)] Loss: 0.001478 Train Epoch: 3 [12160/104000 (12%)] Loss: 0.000351 Train Epoch: 3 [12800/104000 (12%)] Loss: 0.000769 Train Epoch: 3 [13440/104000 (13%)] Loss: 0.000069 Train Epoch: 3 [14080/104000 (14%)] Loss: 0.000273 Train Epoch: 3 [14720/104000 (14%)] Loss: 0.000319 Train Epoch: 3 [15360/104000 (15%)] Loss: 0.000328 Train Epoch: 3 [16000/104000 (15%)] Loss: 0.000319 Train Epoch: 3 [16640/104000 (16%)] Loss: 0.003195 Train Epoch: 3 [17280/104000 (17%)] Loss: 0.020422 Train Epoch: 3 [17920/104000 (17%)] Loss: 0.001275 Train Epoch: 3 [18560/104000 (18%)] Loss: 0.000046 Train Epoch: 3 [19200/104000 (18%)] Loss: 0.000498 Train Epoch: 3 [19840/104000 (19%)] Loss: 0.000359 Train Epoch: 3 [20480/104000 (20%)] Loss: 0.000036 Train Epoch: 3 [21120/104000 (20%)] Loss: 0.000650 Train Epoch: 3 [21760/104000 (21%)] Loss: 0.000145 Train Epoch: 3 [22400/104000 (22%)] Loss: 0.000065 Train Epoch: 3 [23040/104000 (22%)] Loss: 0.000142 Train Epoch: 3 [23680/104000 (23%)] Loss: 0.000053 Train Epoch: 3 [24320/104000 (23%)] Loss: 0.000696 Train Epoch: 3 [24960/104000 (24%)] Loss: 0.000080 Train Epoch: 3 [25600/104000 (25%)] Loss: 0.000205 Train Epoch: 3 [26240/104000 (25%)] Loss: 0.000185 Train Epoch: 3 [26880/104000 (26%)] Loss: 0.000382 Train Epoch: 3 [27520/104000 (26%)] Loss: 0.000008 Train Epoch: 3 [28160/104000 (27%)] Loss: 0.000297 Train Epoch: 3 [28800/104000 (28%)] Loss: 0.000176 Train Epoch: 3 [29440/104000 (28%)] Loss: 0.000226 Train Epoch: 3 [30080/104000 (29%)] Loss: 0.000032 Train Epoch: 3 [30720/104000 (30%)] Loss: 0.000534 Train Epoch: 3 [31360/104000 (30%)] Loss: 0.000174 Train Epoch: 3 [32000/104000 (31%)] Loss: 0.000824 Train Epoch: 3 [32640/104000 (31%)] Loss: 0.000051 Train Epoch: 3 [33280/104000 (32%)] Loss: 0.000113 Train Epoch: 3 [33920/104000 (33%)] Loss: 0.001166 Train Epoch: 3 [34560/104000 (33%)] Loss: 0.000498 Train Epoch: 3 [35200/104000 (34%)] Loss: 0.000053 Train Epoch: 3 [35840/104000 (34%)] Loss: 0.000401 Train Epoch: 3 [36480/104000 (35%)] Loss: 0.000713 Train Epoch: 3 [37120/104000 (36%)] Loss: 0.000182 Train Epoch: 3 [37760/104000 (36%)] Loss: 0.000007 Train Epoch: 3 [38400/104000 (37%)] Loss: 0.000010 Train Epoch: 3 [39040/104000 (38%)] Loss: 0.000439 Train Epoch: 3 [39680/104000 (38%)] Loss: 0.000041 Train Epoch: 3 [40320/104000 (39%)] Loss: 0.000068 Train Epoch: 3 [40960/104000 (39%)] Loss: 0.000030 Train Epoch: 3 [41600/104000 (40%)] Loss: 0.000021 Train Epoch: 3 [42240/104000 (41%)] Loss: 0.000017 Train Epoch: 3 [42880/104000 (41%)] Loss: 0.000019 Train Epoch: 3 [43520/104000 (42%)] Loss: 0.000061 Train Epoch: 3 [44160/104000 (42%)] Loss: 0.000012 Train Epoch: 3 [44800/104000 (43%)] Loss: 0.000309 Train Epoch: 3 [45440/104000 (44%)] Loss: 0.000012 Train Epoch: 3 [46080/104000 (44%)] Loss: 0.000395 Train Epoch: 3 [46720/104000 (45%)] Loss: 0.000037 Train Epoch: 3 [47360/104000 (46%)] Loss: 0.000115 Train Epoch: 3 [48000/104000 (46%)] Loss: 0.000219 Train Epoch: 3 [48640/104000 (47%)] Loss: 0.000003 Train Epoch: 3 [49280/104000 (47%)] Loss: 0.000043 Train Epoch: 3 [49920/104000 (48%)] Loss: 0.000033 Train Epoch: 3 [50560/104000 (49%)] Loss: 0.000105 Train Epoch: 3 [51200/104000 (49%)] Loss: 0.000077 Train Epoch: 3 [51840/104000 (50%)] Loss: 0.000032 Train Epoch: 3 [52480/104000 (50%)] Loss: 0.000009 Train Epoch: 3 [53120/104000 (51%)] Loss: 0.000001 Train Epoch: 3 [53760/104000 (52%)] Loss: 0.000005 Train Epoch: 3 [54400/104000 (52%)] Loss: 0.000008 Train Epoch: 3 [55040/104000 (53%)] Loss: 0.000023 Train Epoch: 3 [55680/104000 (54%)] Loss: 0.000157 Train Epoch: 3 [56320/104000 (54%)] Loss: 0.000024 Train Epoch: 3 [56960/104000 (55%)] Loss: 0.000004 Train Epoch: 3 [57600/104000 (55%)] Loss: 0.000002 Train Epoch: 3 [58240/104000 (56%)] Loss: 0.000016 Train Epoch: 3 [58880/104000 (57%)] Loss: 0.000006 Train Epoch: 3 [59520/104000 (57%)] Loss: 0.000004 Train Epoch: 3 [60160/104000 (58%)] Loss: 0.000003 Train Epoch: 3 [60800/104000 (58%)] Loss: 0.000039 Train Epoch: 3 [61440/104000 (59%)] Loss: 0.009346 Train Epoch: 3 [62080/104000 (60%)] Loss: 0.000003 Train Epoch: 3 [62720/104000 (60%)] Loss: 0.000013 Train Epoch: 3 [63360/104000 (61%)] Loss: 0.000208 Train Epoch: 3 [64000/104000 (62%)] Loss: 0.000003 Train Epoch: 3 [64640/104000 (62%)] Loss: 0.000024 Train Epoch: 3 [65280/104000 (63%)] Loss: 0.000003 Train Epoch: 3 [65920/104000 (63%)] Loss: 0.000018 Train Epoch: 3 [66560/104000 (64%)] Loss: 0.000025 Train Epoch: 3 [67200/104000 (65%)] Loss: 0.000717 Train Epoch: 3 [67840/104000 (65%)] Loss: 0.000052 Train Epoch: 3 [68480/104000 (66%)] Loss: 0.000007 Train Epoch: 3 [69120/104000 (66%)] Loss: 0.000030 Train Epoch: 3 [69760/104000 (67%)] Loss: 0.000991 Train Epoch: 3 [70400/104000 (68%)] Loss: 0.000052 Train Epoch: 3 [71040/104000 (68%)] Loss: 0.000017 Train Epoch: 3 [71680/104000 (69%)] Loss: 0.000007 Train Epoch: 3 [72320/104000 (70%)] Loss: 0.000091 Train Epoch: 3 [72960/104000 (70%)] Loss: 0.000005 Train Epoch: 3 [73600/104000 (71%)] Loss: 0.000018 Train Epoch: 3 [74240/104000 (71%)] Loss: 0.000002 Train Epoch: 3 [74880/104000 (72%)] Loss: 0.000006 Train Epoch: 3 [75520/104000 (73%)] Loss: 0.000011 Train Epoch: 3 [76160/104000 (73%)] Loss: 0.000001 Train Epoch: 3 [76800/104000 (74%)] Loss: 0.000004 Train Epoch: 3 [77440/104000 (74%)] Loss: 0.000010 Train Epoch: 3 [78080/104000 (75%)] Loss: 0.000007 Train Epoch: 3 [78720/104000 (76%)] Loss: 0.000012 Train Epoch: 3 [79360/104000 (76%)] Loss: 0.000001 Train Epoch: 3 [80000/104000 (77%)] Loss: 0.000004 Train Epoch: 3 [80640/104000 (78%)] Loss: 0.000002 Train Epoch: 3 [81280/104000 (78%)] Loss: 0.000075 Train Epoch: 3 [81920/104000 (79%)] Loss: 0.000005 Train Epoch: 3 [82560/104000 (79%)] Loss: 0.000022 Train Epoch: 3 [83200/104000 (80%)] Loss: 0.000009 Train Epoch: 3 [83840/104000 (81%)] Loss: 0.000004 Train Epoch: 3 [84480/104000 (81%)] Loss: 0.000002 Train Epoch: 3 [85120/104000 (82%)] Loss: 0.000005 Train Epoch: 3 [85760/104000 (82%)] Loss: 0.000133 Train Epoch: 3 [86400/104000 (83%)] Loss: 0.000140 Train Epoch: 3 [87040/104000 (84%)] Loss: 0.000006 Train Epoch: 3 [87680/104000 (84%)] Loss: 0.000036 Train Epoch: 3 [88320/104000 (85%)] Loss: 0.000010 Train Epoch: 3 [88960/104000 (86%)] Loss: 0.000005 Train Epoch: 3 [89600/104000 (86%)] Loss: 0.000006 Train Epoch: 3 [90240/104000 (87%)] Loss: 0.000026 Train Epoch: 3 [90880/104000 (87%)] Loss: 0.000003 Train Epoch: 3 [91520/104000 (88%)] Loss: 0.000098 Train Epoch: 3 [92160/104000 (89%)] Loss: 0.000007 Train Epoch: 3 [92800/104000 (89%)] Loss: 0.000002 Train Epoch: 3 [93440/104000 (90%)] Loss: 0.000005 Train Epoch: 3 [94080/104000 (90%)] Loss: 0.000001 Train Epoch: 3 [94720/104000 (91%)] Loss: 0.000005 Train Epoch: 3 [95360/104000 (92%)] Loss: 0.000036 Train Epoch: 3 [96000/104000 (92%)] Loss: 0.000001 Train Epoch: 3 [96640/104000 (93%)] Loss: 0.000001 Train Epoch: 3 [97280/104000 (94%)] Loss: 0.000008 Train Epoch: 3 [97920/104000 (94%)] Loss: 0.000013 Train Epoch: 3 [98560/104000 (95%)] Loss: 0.000003 Train Epoch: 3 [99200/104000 (95%)] Loss: 0.000030 Train Epoch: 3 [99840/104000 (96%)] Loss: 0.000005 Train Epoch: 3 [100480/104000 (97%)] Loss: 0.000014 Train Epoch: 3 [101120/104000 (97%)] Loss: 0.000025 Train Epoch: 3 [101760/104000 (98%)] Loss: 0.000002 Train Epoch: 3 [102400/104000 (98%)] Loss: 0.000003 Train Epoch: 3 [103040/104000 (99%)] Loss: 0.000034 Train Epoch: 3 [103680/104000 (100%)] Loss: 0.000004 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.3871, Accuracy: 25795/30000 (86%) Train Epoch: 4 [0/104000 (0%)] Loss: 0.000001 Train Epoch: 4 [640/104000 (1%)] Loss: 0.000017 Train Epoch: 4 [1280/104000 (1%)] Loss: 0.000013 Train Epoch: 4 [1920/104000 (2%)] Loss: 0.000007 Train Epoch: 4 [2560/104000 (2%)] Loss: 0.000003 Train Epoch: 4 [3200/104000 (3%)] Loss: 0.000014 Train Epoch: 4 [3840/104000 (4%)] Loss: 0.000000 Train Epoch: 4 [4480/104000 (4%)] Loss: 0.000014 Train Epoch: 4 [5120/104000 (5%)] Loss: 0.000003 Train Epoch: 4 [5760/104000 (6%)] Loss: 0.000009 Train Epoch: 4 [6400/104000 (6%)] Loss: 0.000011 Train Epoch: 4 [7040/104000 (7%)] Loss: 0.000000 Train Epoch: 4 [7680/104000 (7%)] Loss: 0.000046 Train Epoch: 4 [8320/104000 (8%)] Loss: 0.000006 Train Epoch: 4 [8960/104000 (9%)] Loss: 0.000004 Train Epoch: 4 [9600/104000 (9%)] Loss: 0.000064 Train Epoch: 4 [10240/104000 (10%)] Loss: 0.000001 Train Epoch: 4 [10880/104000 (10%)] Loss: 0.000859 Train Epoch: 4 [11520/104000 (11%)] Loss: 0.000001 Train Epoch: 4 [12160/104000 (12%)] Loss: 0.000021 Train Epoch: 4 [12800/104000 (12%)] Loss: 0.000045 Train Epoch: 4 [13440/104000 (13%)] Loss: 0.000004 Train Epoch: 4 [14080/104000 (14%)] Loss: 0.000212 Train Epoch: 4 [14720/104000 (14%)] Loss: 0.000001 Train Epoch: 4 [15360/104000 (15%)] Loss: 0.000002 Train Epoch: 4 [16000/104000 (15%)] Loss: 0.000001 Train Epoch: 4 [16640/104000 (16%)] Loss: 0.000008 Train Epoch: 4 [17280/104000 (17%)] Loss: 0.000005 Train Epoch: 4 [17920/104000 (17%)] Loss: 0.000027 Train Epoch: 4 [18560/104000 (18%)] Loss: 0.000003 Train Epoch: 4 [19200/104000 (18%)] Loss: 0.000002 Train Epoch: 4 [19840/104000 (19%)] Loss: 0.000001 Train Epoch: 4 [20480/104000 (20%)] Loss: 0.000003 Train Epoch: 4 [21120/104000 (20%)] Loss: 0.000002 Train Epoch: 4 [21760/104000 (21%)] Loss: 0.000001 Train Epoch: 4 [22400/104000 (22%)] Loss: 0.000008 Train Epoch: 4 [23040/104000 (22%)] Loss: 0.000005 Train Epoch: 4 [23680/104000 (23%)] Loss: 0.000005 Train Epoch: 4 [24320/104000 (23%)] Loss: 0.000003 Train Epoch: 4 [24960/104000 (24%)] Loss: 0.000004 Train Epoch: 4 [25600/104000 (25%)] Loss: 0.000001 Train Epoch: 4 [26240/104000 (25%)] Loss: 0.000001 Train Epoch: 4 [26880/104000 (26%)] Loss: 0.000003 Train Epoch: 4 [27520/104000 (26%)] Loss: 0.000001 Train Epoch: 4 [28160/104000 (27%)] Loss: 0.000083 Train Epoch: 4 [28800/104000 (28%)] Loss: 0.000001 Train Epoch: 4 [29440/104000 (28%)] Loss: 0.000563 Train Epoch: 4 [30080/104000 (29%)] Loss: 0.000001 Train Epoch: 4 [30720/104000 (30%)] Loss: 0.000002 Train Epoch: 4 [31360/104000 (30%)] Loss: 0.000005 Train Epoch: 4 [32000/104000 (31%)] Loss: 0.000008 Train Epoch: 4 [32640/104000 (31%)] Loss: 0.000002 Train Epoch: 4 [33280/104000 (32%)] Loss: 0.000008 Train Epoch: 4 [33920/104000 (33%)] Loss: 0.000001 Train Epoch: 4 [34560/104000 (33%)] Loss: 0.000133 Train Epoch: 4 [35200/104000 (34%)] Loss: 0.000001 Train Epoch: 4 [35840/104000 (34%)] Loss: 0.000011 Train Epoch: 4 [36480/104000 (35%)] Loss: 0.000002 Train Epoch: 4 [37120/104000 (36%)] Loss: 0.000004 Train Epoch: 4 [37760/104000 (36%)] Loss: 0.000001 Train Epoch: 4 [38400/104000 (37%)] Loss: 0.000270 Train Epoch: 4 [39040/104000 (38%)] Loss: 0.000004 Train Epoch: 4 [39680/104000 (38%)] Loss: 0.000012 Train Epoch: 4 [40320/104000 (39%)] Loss: 0.000037 Train Epoch: 4 [40960/104000 (39%)] Loss: 0.000005 Train Epoch: 4 [41600/104000 (40%)] Loss: 0.000005 Train Epoch: 4 [42240/104000 (41%)] Loss: 0.000011 Train Epoch: 4 [42880/104000 (41%)] Loss: 0.000001 Train Epoch: 4 [43520/104000 (42%)] Loss: 0.000015 Train Epoch: 4 [44160/104000 (42%)] Loss: 0.000001 Train Epoch: 4 [44800/104000 (43%)] Loss: 0.000009 Train Epoch: 4 [45440/104000 (44%)] Loss: 0.000007 Train Epoch: 4 [46080/104000 (44%)] Loss: 0.000002 Train Epoch: 4 [46720/104000 (45%)] Loss: 0.000009 Train Epoch: 4 [47360/104000 (46%)] Loss: 0.000006 Train Epoch: 4 [48000/104000 (46%)] Loss: 0.000002 Train Epoch: 4 [48640/104000 (47%)] Loss: 0.000010 Train Epoch: 4 [49280/104000 (47%)] Loss: 0.000001 Train Epoch: 4 [49920/104000 (48%)] Loss: 0.000002 Train Epoch: 4 [50560/104000 (49%)] Loss: 0.000003 Train Epoch: 4 [51200/104000 (49%)] Loss: 0.000001 Train Epoch: 4 [51840/104000 (50%)] Loss: 0.000008 Train Epoch: 4 [52480/104000 (50%)] Loss: 0.000012 Train Epoch: 4 [53120/104000 (51%)] Loss: 0.000158 Train Epoch: 4 [53760/104000 (52%)] Loss: 0.000009 Train Epoch: 4 [54400/104000 (52%)] Loss: 0.000003 Train Epoch: 4 [55040/104000 (53%)] Loss: 0.000001 Train Epoch: 4 [55680/104000 (54%)] Loss: 0.000001 Train Epoch: 4 [56320/104000 (54%)] Loss: 0.000002 Train Epoch: 4 [56960/104000 (55%)] Loss: 0.000002 Train Epoch: 4 [57600/104000 (55%)] Loss: 0.000003 Train Epoch: 4 [58240/104000 (56%)] Loss: 0.000008 Train Epoch: 4 [58880/104000 (57%)] Loss: 0.000008 Train Epoch: 4 [59520/104000 (57%)] Loss: 0.000013 Train Epoch: 4 [60160/104000 (58%)] Loss: 0.000002 Train Epoch: 4 [60800/104000 (58%)] Loss: 0.000002 Train Epoch: 4 [61440/104000 (59%)] Loss: 0.000019 Train Epoch: 4 [62080/104000 (60%)] Loss: 0.000004 Train Epoch: 4 [62720/104000 (60%)] Loss: 0.000002 Train Epoch: 4 [63360/104000 (61%)] Loss: 0.000003 Train Epoch: 4 [64000/104000 (62%)] Loss: 0.000003 Train Epoch: 4 [64640/104000 (62%)] Loss: 0.000004 Train Epoch: 4 [65280/104000 (63%)] Loss: 0.000002 Train Epoch: 4 [65920/104000 (63%)] Loss: 0.000004 Train Epoch: 4 [66560/104000 (64%)] Loss: 0.000001 Train Epoch: 4 [67200/104000 (65%)] Loss: 0.000007 Train Epoch: 4 [67840/104000 (65%)] Loss: 0.000030 Train Epoch: 4 [68480/104000 (66%)] Loss: 0.000031 Train Epoch: 4 [69120/104000 (66%)] Loss: 0.000087 Train Epoch: 4 [69760/104000 (67%)] Loss: 0.000059 Train Epoch: 4 [70400/104000 (68%)] Loss: 0.000084 Train Epoch: 4 [71040/104000 (68%)] Loss: 0.000006 Train Epoch: 4 [71680/104000 (69%)] Loss: 0.000003 Train Epoch: 4 [72320/104000 (70%)] Loss: 0.000064 Train Epoch: 4 [72960/104000 (70%)] Loss: 0.000001 Train Epoch: 4 [73600/104000 (71%)] Loss: 0.000004 Train Epoch: 4 [74240/104000 (71%)] Loss: 0.000004 Train Epoch: 4 [74880/104000 (72%)] Loss: 0.000000 Train Epoch: 4 [75520/104000 (73%)] Loss: 0.000001 Train Epoch: 4 [76160/104000 (73%)] Loss: 0.000034 Train Epoch: 4 [76800/104000 (74%)] Loss: 0.000062 Train Epoch: 4 [77440/104000 (74%)] Loss: 0.000008 Train Epoch: 4 [78080/104000 (75%)] Loss: 0.000015 Train Epoch: 4 [78720/104000 (76%)] Loss: 0.000000 Train Epoch: 4 [79360/104000 (76%)] Loss: 0.000002 Train Epoch: 4 [80000/104000 (77%)] Loss: 0.000001 Train Epoch: 4 [80640/104000 (78%)] Loss: 0.000000 Train Epoch: 4 [81280/104000 (78%)] Loss: 0.000008 Train Epoch: 4 [81920/104000 (79%)] Loss: 0.000007 Train Epoch: 4 [82560/104000 (79%)] Loss: 0.000003 Train Epoch: 4 [83200/104000 (80%)] Loss: 0.000015 Train Epoch: 4 [83840/104000 (81%)] Loss: 0.000003 Train Epoch: 4 [84480/104000 (81%)] Loss: 0.000007 Train Epoch: 4 [85120/104000 (82%)] Loss: 0.000001 Train Epoch: 4 [85760/104000 (82%)] Loss: 0.000002 Train Epoch: 4 [86400/104000 (83%)] Loss: 0.000004 Train Epoch: 4 [87040/104000 (84%)] Loss: 0.000227 Train Epoch: 4 [87680/104000 (84%)] Loss: 0.000003 Train Epoch: 4 [88320/104000 (85%)] Loss: 0.000003 Train Epoch: 4 [88960/104000 (86%)] Loss: 0.000005 Train Epoch: 4 [89600/104000 (86%)] Loss: 0.000004 Train Epoch: 4 [90240/104000 (87%)] Loss: 0.000001 Train Epoch: 4 [90880/104000 (87%)] Loss: 0.000002 Train Epoch: 4 [91520/104000 (88%)] Loss: 0.000001 Train Epoch: 4 [92160/104000 (89%)] Loss: 0.000001 Train Epoch: 4 [92800/104000 (89%)] Loss: 0.000053 Train Epoch: 4 [93440/104000 (90%)] Loss: 0.000002 Train Epoch: 4 [94080/104000 (90%)] Loss: 0.000013 Train Epoch: 4 [94720/104000 (91%)] Loss: 0.000020 Train Epoch: 4 [95360/104000 (92%)] Loss: 0.000001 Train Epoch: 4 [96000/104000 (92%)] Loss: 0.000009 Train Epoch: 4 [96640/104000 (93%)] Loss: 0.000001 Train Epoch: 4 [97280/104000 (94%)] Loss: 0.000074 Train Epoch: 4 [97920/104000 (94%)] Loss: 0.000001 Train Epoch: 4 [98560/104000 (95%)] Loss: 0.000016 Train Epoch: 4 [99200/104000 (95%)] Loss: 0.000002 Train Epoch: 4 [99840/104000 (96%)] Loss: 0.000001 Train Epoch: 4 [100480/104000 (97%)] Loss: 0.000003 Train Epoch: 4 [101120/104000 (97%)] Loss: 0.000004 Train Epoch: 4 [101760/104000 (98%)] Loss: 0.000006 Train Epoch: 4 [102400/104000 (98%)] Loss: 0.000001 Train Epoch: 4 [103040/104000 (99%)] Loss: 0.000000 Train Epoch: 4 [103680/104000 (100%)] Loss: 0.000006 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.3498, Accuracy: 25909/30000 (86%) Train Epoch: 5 [0/104000 (0%)] Loss: 0.000037 Train Epoch: 5 [640/104000 (1%)] Loss: 0.000001 Train Epoch: 5 [1280/104000 (1%)] Loss: 0.000001 Train Epoch: 5 [1920/104000 (2%)] Loss: 0.000003 Train Epoch: 5 [2560/104000 (2%)] Loss: 0.000002 Train Epoch: 5 [3200/104000 (3%)] Loss: 0.000003 Train Epoch: 5 [3840/104000 (4%)] Loss: 0.000001 Train Epoch: 5 [4480/104000 (4%)] Loss: 0.000009 Train Epoch: 5 [5120/104000 (5%)] Loss: 0.000000 Train Epoch: 5 [5760/104000 (6%)] Loss: 0.000002 Train Epoch: 5 [6400/104000 (6%)] Loss: 0.000002 Train Epoch: 5 [7040/104000 (7%)] Loss: 0.000004 Train Epoch: 5 [7680/104000 (7%)] Loss: 0.000001 Train Epoch: 5 [8320/104000 (8%)] Loss: 0.000001 Train Epoch: 5 [8960/104000 (9%)] Loss: 0.000002 Train Epoch: 5 [9600/104000 (9%)] Loss: 0.000003 Train Epoch: 5 [10240/104000 (10%)] Loss: 0.000002 Train Epoch: 5 [10880/104000 (10%)] Loss: 0.000000 Train Epoch: 5 [11520/104000 (11%)] Loss: 0.000002 Train Epoch: 5 [12160/104000 (12%)] Loss: 0.000158 Train Epoch: 5 [12800/104000 (12%)] Loss: 0.000001 Train Epoch: 5 [13440/104000 (13%)] Loss: 0.000003 Train Epoch: 5 [14080/104000 (14%)] Loss: 0.000004 Train Epoch: 5 [14720/104000 (14%)] Loss: 0.000007 Train Epoch: 5 [15360/104000 (15%)] Loss: 0.000001 Train Epoch: 5 [16000/104000 (15%)] Loss: 0.000001 Train Epoch: 5 [16640/104000 (16%)] Loss: 0.000001 Train Epoch: 5 [17280/104000 (17%)] Loss: 0.000002 Train Epoch: 5 [17920/104000 (17%)] Loss: 0.000001 Train Epoch: 5 [18560/104000 (18%)] Loss: 0.000005 Train Epoch: 5 [19200/104000 (18%)] Loss: 0.000001 Train Epoch: 5 [19840/104000 (19%)] Loss: 0.000003 Train Epoch: 5 [20480/104000 (20%)] Loss: 0.000004 Train Epoch: 5 [21120/104000 (20%)] Loss: 0.000001 Train Epoch: 5 [21760/104000 (21%)] Loss: 0.000001 Train Epoch: 5 [22400/104000 (22%)] Loss: 0.000002 Train Epoch: 5 [23040/104000 (22%)] Loss: 0.000003 Train Epoch: 5 [23680/104000 (23%)] Loss: 0.000002 Train Epoch: 5 [24320/104000 (23%)] Loss: 0.000052 Train Epoch: 5 [24960/104000 (24%)] Loss: 0.000039 Train Epoch: 5 [25600/104000 (25%)] Loss: 0.000011 Train Epoch: 5 [26240/104000 (25%)] Loss: 0.000009 Train Epoch: 5 [26880/104000 (26%)] Loss: 0.000010 Train Epoch: 5 [27520/104000 (26%)] Loss: 0.000046 Train Epoch: 5 [28160/104000 (27%)] Loss: 0.000002 Train Epoch: 5 [28800/104000 (28%)] Loss: 0.000004 Train Epoch: 5 [29440/104000 (28%)] Loss: 0.000001 Train Epoch: 5 [30080/104000 (29%)] Loss: 0.000001 Train Epoch: 5 [30720/104000 (30%)] Loss: 0.000042 Train Epoch: 5 [31360/104000 (30%)] Loss: 0.000001 Train Epoch: 5 [32000/104000 (31%)] Loss: 0.000003 Train Epoch: 5 [32640/104000 (31%)] Loss: 0.000012 Train Epoch: 5 [33280/104000 (32%)] Loss: 0.000004 Train Epoch: 5 [33920/104000 (33%)] Loss: 0.000002 Train Epoch: 5 [34560/104000 (33%)] Loss: 0.000000 Train Epoch: 5 [35200/104000 (34%)] Loss: 0.000011 Train Epoch: 5 [35840/104000 (34%)] Loss: 0.000001 Train Epoch: 5 [36480/104000 (35%)] Loss: 0.000000 Train Epoch: 5 [37120/104000 (36%)] Loss: 0.000679 Train Epoch: 5 [37760/104000 (36%)] Loss: 0.000003 Train Epoch: 5 [38400/104000 (37%)] Loss: 0.000006 Train Epoch: 5 [39040/104000 (38%)] Loss: 0.000004 Train Epoch: 5 [39680/104000 (38%)] Loss: 0.000001 Train Epoch: 5 [40320/104000 (39%)] Loss: 0.000001 Train Epoch: 5 [40960/104000 (39%)] Loss: 0.000001 Train Epoch: 5 [41600/104000 (40%)] Loss: 0.000000 Train Epoch: 5 [42240/104000 (41%)] Loss: 0.000044 Train Epoch: 5 [42880/104000 (41%)] Loss: 0.000001 Train Epoch: 5 [43520/104000 (42%)] Loss: 0.000002 Train Epoch: 5 [44160/104000 (42%)] Loss: 0.000007 Train Epoch: 5 [44800/104000 (43%)] Loss: 0.000002 Train Epoch: 5 [45440/104000 (44%)] Loss: 0.000005 Train Epoch: 5 [46080/104000 (44%)] Loss: 0.000006 Train Epoch: 5 [46720/104000 (45%)] Loss: 0.000011 Train Epoch: 5 [47360/104000 (46%)] Loss: 0.000000 Train Epoch: 5 [48000/104000 (46%)] Loss: 0.000001 Train Epoch: 5 [48640/104000 (47%)] Loss: 0.000001 Train Epoch: 5 [49280/104000 (47%)] Loss: 0.000013 Train Epoch: 5 [49920/104000 (48%)] Loss: 0.000007 Train Epoch: 5 [50560/104000 (49%)] Loss: 0.000001 Train Epoch: 5 [51200/104000 (49%)] Loss: 0.000008 Train Epoch: 5 [51840/104000 (50%)] Loss: 0.000000 Train Epoch: 5 [52480/104000 (50%)] Loss: 0.000003 Train Epoch: 5 [53120/104000 (51%)] Loss: 0.000002 Train Epoch: 5 [53760/104000 (52%)] Loss: 0.000004 Train Epoch: 5 [54400/104000 (52%)] Loss: 0.000007 Train Epoch: 5 [55040/104000 (53%)] Loss: 0.000002 Train Epoch: 5 [55680/104000 (54%)] Loss: 0.000001 Train Epoch: 5 [56320/104000 (54%)] Loss: 0.000000 Train Epoch: 5 [56960/104000 (55%)] Loss: 0.000001 Train Epoch: 5 [57600/104000 (55%)] Loss: 0.000002 Train Epoch: 5 [58240/104000 (56%)] Loss: 0.000004 Train Epoch: 5 [58880/104000 (57%)] Loss: 0.000007 Train Epoch: 5 [59520/104000 (57%)] Loss: 0.000001 Train Epoch: 5 [60160/104000 (58%)] Loss: 0.000002 Train Epoch: 5 [60800/104000 (58%)] Loss: 0.000002 Train Epoch: 5 [61440/104000 (59%)] Loss: 0.000192 Train Epoch: 5 [62080/104000 (60%)] Loss: 0.000000 Train Epoch: 5 [62720/104000 (60%)] Loss: 0.000002 Train Epoch: 5 [63360/104000 (61%)] Loss: 0.000002 Train Epoch: 5 [64000/104000 (62%)] Loss: 0.000004 Train Epoch: 5 [64640/104000 (62%)] Loss: 0.000029 Train Epoch: 5 [65280/104000 (63%)] Loss: 0.000001 Train Epoch: 5 [65920/104000 (63%)] Loss: 0.000000 Train Epoch: 5 [66560/104000 (64%)] Loss: 0.000004 Train Epoch: 5 [67200/104000 (65%)] Loss: 0.000002 Train Epoch: 5 [67840/104000 (65%)] Loss: 0.000031 Train Epoch: 5 [68480/104000 (66%)] Loss: 0.000002 Train Epoch: 5 [69120/104000 (66%)] Loss: 0.000003 Train Epoch: 5 [69760/104000 (67%)] Loss: 0.000001 Train Epoch: 5 [70400/104000 (68%)] Loss: 0.000001 Train Epoch: 5 [71040/104000 (68%)] Loss: 0.000006 Train Epoch: 5 [71680/104000 (69%)] Loss: 0.000002 Train Epoch: 5 [72320/104000 (70%)] Loss: 0.000006 Train Epoch: 5 [72960/104000 (70%)] Loss: 0.000001 Train Epoch: 5 [73600/104000 (71%)] Loss: 0.000001 Train Epoch: 5 [74240/104000 (71%)] Loss: 0.000004 Train Epoch: 5 [74880/104000 (72%)] Loss: 0.000001 Train Epoch: 5 [75520/104000 (73%)] Loss: 0.000001 Train Epoch: 5 [76160/104000 (73%)] Loss: 0.000003 Train Epoch: 5 [76800/104000 (74%)] Loss: 0.000061 Train Epoch: 5 [77440/104000 (74%)] Loss: 0.000007 Train Epoch: 5 [78080/104000 (75%)] Loss: 0.000001 Train Epoch: 5 [78720/104000 (76%)] Loss: 0.000005 Train Epoch: 5 [79360/104000 (76%)] Loss: 0.000003 Train Epoch: 5 [80000/104000 (77%)] Loss: 0.000009 Train Epoch: 5 [80640/104000 (78%)] Loss: 0.000017 Train Epoch: 5 [81280/104000 (78%)] Loss: 0.000051 Train Epoch: 5 [81920/104000 (79%)] Loss: 0.000001 Train Epoch: 5 [82560/104000 (79%)] Loss: 0.000002 Train Epoch: 5 [83200/104000 (80%)] Loss: 0.000005 Train Epoch: 5 [83840/104000 (81%)] Loss: 0.000006 Train Epoch: 5 [84480/104000 (81%)] Loss: 0.000007 Train Epoch: 5 [85120/104000 (82%)] Loss: 0.000006 Train Epoch: 5 [85760/104000 (82%)] Loss: 0.000060 Train Epoch: 5 [86400/104000 (83%)] Loss: 0.000002 Train Epoch: 5 [87040/104000 (84%)] Loss: 0.000003 Train Epoch: 5 [87680/104000 (84%)] Loss: 0.000021 Train Epoch: 5 [88320/104000 (85%)] Loss: 0.000002 Train Epoch: 5 [88960/104000 (86%)] Loss: 0.000003 Train Epoch: 5 [89600/104000 (86%)] Loss: 0.000031 Train Epoch: 5 [90240/104000 (87%)] Loss: 0.000000 Train Epoch: 5 [90880/104000 (87%)] Loss: 0.000003 Train Epoch: 5 [91520/104000 (88%)] Loss: 0.000001 Train Epoch: 5 [92160/104000 (89%)] Loss: 0.000001 Train Epoch: 5 [92800/104000 (89%)] Loss: 0.000001 Train Epoch: 5 [93440/104000 (90%)] Loss: 0.000002 Train Epoch: 5 [94080/104000 (90%)] Loss: 0.000001 Train Epoch: 5 [94720/104000 (91%)] Loss: 0.000001 Train Epoch: 5 [95360/104000 (92%)] Loss: 0.000001 Train Epoch: 5 [96000/104000 (92%)] Loss: 0.000006 Train Epoch: 5 [96640/104000 (93%)] Loss: 0.000064 Train Epoch: 5 [97280/104000 (94%)] Loss: 0.000003 Train Epoch: 5 [97920/104000 (94%)] Loss: 0.000001 Train Epoch: 5 [98560/104000 (95%)] Loss: 0.000021 Train Epoch: 5 [99200/104000 (95%)] Loss: 0.000017 Train Epoch: 5 [99840/104000 (96%)] Loss: 0.000006 Train Epoch: 5 [100480/104000 (97%)] Loss: 0.000000 Train Epoch: 5 [101120/104000 (97%)] Loss: 0.000002 Train Epoch: 5 [101760/104000 (98%)] Loss: 0.000014 Train Epoch: 5 [102400/104000 (98%)] Loss: 0.000002 Train Epoch: 5 [103040/104000 (99%)] Loss: 0.000006 Train Epoch: 5 [103680/104000 (100%)] Loss: 0.000003 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2745, Accuracy: 26568/30000 (89%) [100.0] [88.56] The number of neurons in CNN layer is 5 Train Epoch: 1 [0/104000 (0%)] Loss: 0.735635 Train Epoch: 1 [640/104000 (1%)] Loss: 0.006729 Train Epoch: 1 [1280/104000 (1%)] Loss: 0.008744 Train Epoch: 1 [1920/104000 (2%)] Loss: 0.003856 Train Epoch: 1 [2560/104000 (2%)] Loss: 0.002027 Train Epoch: 1 [3200/104000 (3%)] Loss: 0.001791 Train Epoch: 1 [3840/104000 (4%)] Loss: 0.001974 Train Epoch: 1 [4480/104000 (4%)] Loss: 0.002818 Train Epoch: 1 [5120/104000 (5%)] Loss: 0.001130 Train Epoch: 1 [5760/104000 (6%)] Loss: 0.001281 Train Epoch: 1 [6400/104000 (6%)] Loss: 0.001820 Train Epoch: 1 [7040/104000 (7%)] Loss: 0.000577 Train Epoch: 1 [7680/104000 (7%)] Loss: 0.002283 Train Epoch: 1 [8320/104000 (8%)] Loss: 0.001235 Train Epoch: 1 [8960/104000 (9%)] Loss: 0.000694 Train Epoch: 1 [9600/104000 (9%)] Loss: 0.001370 Train Epoch: 1 [10240/104000 (10%)] Loss: 0.000621 Train Epoch: 1 [10880/104000 (10%)] Loss: 0.000582 Train Epoch: 1 [11520/104000 (11%)] Loss: 0.002235 Train Epoch: 1 [12160/104000 (12%)] Loss: 0.000568 Train Epoch: 1 [12800/104000 (12%)] Loss: 0.000624 Train Epoch: 1 [13440/104000 (13%)] Loss: 0.000438 Train Epoch: 1 [14080/104000 (14%)] Loss: 0.000685 Train Epoch: 1 [14720/104000 (14%)] Loss: 0.000693 Train Epoch: 1 [15360/104000 (15%)] Loss: 0.000425 Train Epoch: 1 [16000/104000 (15%)] Loss: 0.000955 Train Epoch: 1 [16640/104000 (16%)] Loss: 0.000375 Train Epoch: 1 [17280/104000 (17%)] Loss: 0.001304 Train Epoch: 1 [17920/104000 (17%)] Loss: 0.000319 Train Epoch: 1 [18560/104000 (18%)] Loss: 0.000427 Train Epoch: 1 [19200/104000 (18%)] Loss: 0.000714 Train Epoch: 1 [19840/104000 (19%)] Loss: 0.000396 Train Epoch: 1 [20480/104000 (20%)] Loss: 0.000221 Train Epoch: 1 [21120/104000 (20%)] Loss: 0.001221 Train Epoch: 1 [21760/104000 (21%)] Loss: 0.000625 Train Epoch: 1 [22400/104000 (22%)] Loss: 0.000173 Train Epoch: 1 [23040/104000 (22%)] Loss: 0.000290 Train Epoch: 1 [23680/104000 (23%)] Loss: 0.000173 Train Epoch: 1 [24320/104000 (23%)] Loss: 0.000257 Train Epoch: 1 [24960/104000 (24%)] Loss: 0.000416 Train Epoch: 1 [25600/104000 (25%)] Loss: 0.000329 Train Epoch: 1 [26240/104000 (25%)] Loss: 0.001052 Train Epoch: 1 [26880/104000 (26%)] Loss: 0.000258 Train Epoch: 1 [27520/104000 (26%)] Loss: 0.000182 Train Epoch: 1 [28160/104000 (27%)] Loss: 0.000222 Train Epoch: 1 [28800/104000 (28%)] Loss: 0.001760 Train Epoch: 1 [29440/104000 (28%)] Loss: 0.000554 Train Epoch: 1 [30080/104000 (29%)] Loss: 0.000418 Train Epoch: 1 [30720/104000 (30%)] Loss: 0.000221 Train Epoch: 1 [31360/104000 (30%)] Loss: 0.000220 Train Epoch: 1 [32000/104000 (31%)] Loss: 0.000295 Train Epoch: 1 [32640/104000 (31%)] Loss: 0.000210 Train Epoch: 1 [33280/104000 (32%)] Loss: 0.000116 Train Epoch: 1 [33920/104000 (33%)] Loss: 0.002035 Train Epoch: 1 [34560/104000 (33%)] Loss: 0.000297 Train Epoch: 1 [35200/104000 (34%)] Loss: 0.000158 Train Epoch: 1 [35840/104000 (34%)] Loss: 0.000380 Train Epoch: 1 [36480/104000 (35%)] Loss: 0.000463 Train Epoch: 1 [37120/104000 (36%)] Loss: 0.000155 Train Epoch: 1 [37760/104000 (36%)] Loss: 0.000535 Train Epoch: 1 [38400/104000 (37%)] Loss: 0.000258 Train Epoch: 1 [39040/104000 (38%)] Loss: 0.000307 Train Epoch: 1 [39680/104000 (38%)] Loss: 0.000672 Train Epoch: 1 [40320/104000 (39%)] Loss: 0.000382 Train Epoch: 1 [40960/104000 (39%)] Loss: 0.000181 Train Epoch: 1 [41600/104000 (40%)] Loss: 0.000159 Train Epoch: 1 [42240/104000 (41%)] Loss: 0.000101 Train Epoch: 1 [42880/104000 (41%)] Loss: 0.000263 Train Epoch: 1 [43520/104000 (42%)] Loss: 0.000222 Train Epoch: 1 [44160/104000 (42%)] Loss: 0.000363 Train Epoch: 1 [44800/104000 (43%)] Loss: 0.000188 Train Epoch: 1 [45440/104000 (44%)] Loss: 0.000522 Train Epoch: 1 [46080/104000 (44%)] Loss: 0.000156 Train Epoch: 1 [46720/104000 (45%)] Loss: 0.000356 Train Epoch: 1 [47360/104000 (46%)] Loss: 0.000110 Train Epoch: 1 [48000/104000 (46%)] Loss: 0.000104 Train Epoch: 1 [48640/104000 (47%)] Loss: 0.001513 Train Epoch: 1 [49280/104000 (47%)] Loss: 0.000110 Train Epoch: 1 [49920/104000 (48%)] Loss: 0.000910 Train Epoch: 1 [50560/104000 (49%)] Loss: 0.000097 Train Epoch: 1 [51200/104000 (49%)] Loss: 0.000221 Train Epoch: 1 [51840/104000 (50%)] Loss: 0.000094 Train Epoch: 1 [52480/104000 (50%)] Loss: 0.000126 Train Epoch: 1 [53120/104000 (51%)] Loss: 0.000117 Train Epoch: 1 [53760/104000 (52%)] Loss: 0.000074 Train Epoch: 1 [54400/104000 (52%)] Loss: 0.000081 Train Epoch: 1 [55040/104000 (53%)] Loss: 0.000126 Train Epoch: 1 [55680/104000 (54%)] Loss: 0.000236 Train Epoch: 1 [56320/104000 (54%)] Loss: 0.000757 Train Epoch: 1 [56960/104000 (55%)] Loss: 0.000106 Train Epoch: 1 [57600/104000 (55%)] Loss: 0.000110 Train Epoch: 1 [58240/104000 (56%)] Loss: 0.000213 Train Epoch: 1 [58880/104000 (57%)] Loss: 0.002330 Train Epoch: 1 [59520/104000 (57%)] Loss: 0.000128 Train Epoch: 1 [60160/104000 (58%)] Loss: 0.000242 Train Epoch: 1 [60800/104000 (58%)] Loss: 0.000098 Train Epoch: 1 [61440/104000 (59%)] Loss: 0.000131 Train Epoch: 1 [62080/104000 (60%)] Loss: 0.000151 Train Epoch: 1 [62720/104000 (60%)] Loss: 0.000059 Train Epoch: 1 [63360/104000 (61%)] Loss: 0.000225 Train Epoch: 1 [64000/104000 (62%)] Loss: 0.000086 Train Epoch: 1 [64640/104000 (62%)] Loss: 0.000091 Train Epoch: 1 [65280/104000 (63%)] Loss: 0.000537 Train Epoch: 1 [65920/104000 (63%)] Loss: 0.000686 Train Epoch: 1 [66560/104000 (64%)] Loss: 0.000159 Train Epoch: 1 [67200/104000 (65%)] Loss: 0.000342 Train Epoch: 1 [67840/104000 (65%)] Loss: 0.000073 Train Epoch: 1 [68480/104000 (66%)] Loss: 0.000064 Train Epoch: 1 [69120/104000 (66%)] Loss: 0.000098 Train Epoch: 1 [69760/104000 (67%)] Loss: 0.000095 Train Epoch: 1 [70400/104000 (68%)] Loss: 0.000273 Train Epoch: 1 [71040/104000 (68%)] Loss: 0.000096 Train Epoch: 1 [71680/104000 (69%)] Loss: 0.000077 Train Epoch: 1 [72320/104000 (70%)] Loss: 0.000106 Train Epoch: 1 [72960/104000 (70%)] Loss: 0.000131 Train Epoch: 1 [73600/104000 (71%)] Loss: 0.000170 Train Epoch: 1 [74240/104000 (71%)] Loss: 0.000123 Train Epoch: 1 [74880/104000 (72%)] Loss: 0.000081 Train Epoch: 1 [75520/104000 (73%)] Loss: 0.000088 Train Epoch: 1 [76160/104000 (73%)] Loss: 0.000166 Train Epoch: 1 [76800/104000 (74%)] Loss: 0.000057 Train Epoch: 1 [77440/104000 (74%)] Loss: 0.000100 Train Epoch: 1 [78080/104000 (75%)] Loss: 0.000197 Train Epoch: 1 [78720/104000 (76%)] Loss: 0.000060 Train Epoch: 1 [79360/104000 (76%)] Loss: 0.000106 Train Epoch: 1 [80000/104000 (77%)] Loss: 0.000139 Train Epoch: 1 [80640/104000 (78%)] Loss: 0.000069 Train Epoch: 1 [81280/104000 (78%)] Loss: 0.000104 Train Epoch: 1 [81920/104000 (79%)] Loss: 0.000082 Train Epoch: 1 [82560/104000 (79%)] Loss: 0.000159 Train Epoch: 1 [83200/104000 (80%)] Loss: 0.000061 Train Epoch: 1 [83840/104000 (81%)] Loss: 0.000215 Train Epoch: 1 [84480/104000 (81%)] Loss: 0.000723 Train Epoch: 1 [85120/104000 (82%)] Loss: 0.000050 Train Epoch: 1 [85760/104000 (82%)] Loss: 0.000138 Train Epoch: 1 [86400/104000 (83%)] Loss: 0.000088 Train Epoch: 1 [87040/104000 (84%)] Loss: 0.000291 Train Epoch: 1 [87680/104000 (84%)] Loss: 0.000090 Train Epoch: 1 [88320/104000 (85%)] Loss: 0.000047 Train Epoch: 1 [88960/104000 (86%)] Loss: 0.000045 Train Epoch: 1 [89600/104000 (86%)] Loss: 0.000162 Train Epoch: 1 [90240/104000 (87%)] Loss: 0.000181 Train Epoch: 1 [90880/104000 (87%)] Loss: 0.000334 Train Epoch: 1 [91520/104000 (88%)] Loss: 0.000049 Train Epoch: 1 [92160/104000 (89%)] Loss: 0.000166 Train Epoch: 1 [92800/104000 (89%)] Loss: 0.000054 Train Epoch: 1 [93440/104000 (90%)] Loss: 0.000352 Train Epoch: 1 [94080/104000 (90%)] Loss: 0.000066 Train Epoch: 1 [94720/104000 (91%)] Loss: 0.000109 Train Epoch: 1 [95360/104000 (92%)] Loss: 0.000052 Train Epoch: 1 [96000/104000 (92%)] Loss: 0.000066 Train Epoch: 1 [96640/104000 (93%)] Loss: 0.000053 Train Epoch: 1 [97280/104000 (94%)] Loss: 0.000082 Train Epoch: 1 [97920/104000 (94%)] Loss: 0.000039 Train Epoch: 1 [98560/104000 (95%)] Loss: 0.000065 Train Epoch: 1 [99200/104000 (95%)] Loss: 0.000068 Train Epoch: 1 [99840/104000 (96%)] Loss: 0.000054 Train Epoch: 1 [100480/104000 (97%)] Loss: 0.000082 Train Epoch: 1 [101120/104000 (97%)] Loss: 0.000055 Train Epoch: 1 [101760/104000 (98%)] Loss: 0.000100 Train Epoch: 1 [102400/104000 (98%)] Loss: 0.000185 Train Epoch: 1 [103040/104000 (99%)] Loss: 0.000271 Train Epoch: 1 [103680/104000 (100%)] Loss: 0.000036 Test set: Average loss: 0.0001, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2616, Accuracy: 26099/30000 (87%) Train Epoch: 2 [0/104000 (0%)] Loss: 0.000091 Train Epoch: 2 [640/104000 (1%)] Loss: 0.000039 Train Epoch: 2 [1280/104000 (1%)] Loss: 0.000085 Train Epoch: 2 [1920/104000 (2%)] Loss: 0.000152 Train Epoch: 2 [2560/104000 (2%)] Loss: 0.000086 Train Epoch: 2 [3200/104000 (3%)] Loss: 0.000064 Train Epoch: 2 [3840/104000 (4%)] Loss: 0.000059 Train Epoch: 2 [4480/104000 (4%)] Loss: 0.000649 Train Epoch: 2 [5120/104000 (5%)] Loss: 0.000282 Train Epoch: 2 [5760/104000 (6%)] Loss: 0.000036 Train Epoch: 2 [6400/104000 (6%)] Loss: 0.000047 Train Epoch: 2 [7040/104000 (7%)] Loss: 0.000039 Train Epoch: 2 [7680/104000 (7%)] Loss: 0.000091 Train Epoch: 2 [8320/104000 (8%)] Loss: 0.000032 Train Epoch: 2 [8960/104000 (9%)] Loss: 0.000316 Train Epoch: 2 [9600/104000 (9%)] Loss: 0.000034 Train Epoch: 2 [10240/104000 (10%)] Loss: 0.000067 Train Epoch: 2 [10880/104000 (10%)] Loss: 0.000373 Train Epoch: 2 [11520/104000 (11%)] Loss: 0.000094 Train Epoch: 2 [12160/104000 (12%)] Loss: 0.000035 Train Epoch: 2 [12800/104000 (12%)] Loss: 0.000582 Train Epoch: 2 [13440/104000 (13%)] Loss: 0.000043 Train Epoch: 2 [14080/104000 (14%)] Loss: 0.000256 Train Epoch: 2 [14720/104000 (14%)] Loss: 0.000051 Train Epoch: 2 [15360/104000 (15%)] Loss: 0.000048 Train Epoch: 2 [16000/104000 (15%)] Loss: 0.000040 Train Epoch: 2 [16640/104000 (16%)] Loss: 0.000048 Train Epoch: 2 [17280/104000 (17%)] Loss: 0.000190 Train Epoch: 2 [17920/104000 (17%)] Loss: 0.000038 Train Epoch: 2 [18560/104000 (18%)] Loss: 0.000502 Train Epoch: 2 [19200/104000 (18%)] Loss: 0.000037 Train Epoch: 2 [19840/104000 (19%)] Loss: 0.000034 Train Epoch: 2 [20480/104000 (20%)] Loss: 0.000313 Train Epoch: 2 [21120/104000 (20%)] Loss: 0.000083 Train Epoch: 2 [21760/104000 (21%)] Loss: 0.000061 Train Epoch: 2 [22400/104000 (22%)] Loss: 0.000113 Train Epoch: 2 [23040/104000 (22%)] Loss: 0.000027 Train Epoch: 2 [23680/104000 (23%)] Loss: 0.000078 Train Epoch: 2 [24320/104000 (23%)] Loss: 0.000094 Train Epoch: 2 [24960/104000 (24%)] Loss: 0.000041 Train Epoch: 2 [25600/104000 (25%)] Loss: 0.000130 Train Epoch: 2 [26240/104000 (25%)] Loss: 0.000056 Train Epoch: 2 [26880/104000 (26%)] Loss: 0.002189 Train Epoch: 2 [27520/104000 (26%)] Loss: 0.000205 Train Epoch: 2 [28160/104000 (27%)] Loss: 0.000045 Train Epoch: 2 [28800/104000 (28%)] Loss: 0.000070 Train Epoch: 2 [29440/104000 (28%)] Loss: 0.000073 Train Epoch: 2 [30080/104000 (29%)] Loss: 0.000064 Train Epoch: 2 [30720/104000 (30%)] Loss: 0.000055 Train Epoch: 2 [31360/104000 (30%)] Loss: 0.000139 Train Epoch: 2 [32000/104000 (31%)] Loss: 0.000033 Train Epoch: 2 [32640/104000 (31%)] Loss: 0.000054 Train Epoch: 2 [33280/104000 (32%)] Loss: 0.000057 Train Epoch: 2 [33920/104000 (33%)] Loss: 0.000062 Train Epoch: 2 [34560/104000 (33%)] Loss: 0.000061 Train Epoch: 2 [35200/104000 (34%)] Loss: 0.000050 Train Epoch: 2 [35840/104000 (34%)] Loss: 0.000143 Train Epoch: 2 [36480/104000 (35%)] Loss: 0.000057 Train Epoch: 2 [37120/104000 (36%)] Loss: 0.000041 Train Epoch: 2 [37760/104000 (36%)] Loss: 0.000286 Train Epoch: 2 [38400/104000 (37%)] Loss: 0.000020 Train Epoch: 2 [39040/104000 (38%)] Loss: 0.000036 Train Epoch: 2 [39680/104000 (38%)] Loss: 0.000020 Train Epoch: 2 [40320/104000 (39%)] Loss: 0.000025 Train Epoch: 2 [40960/104000 (39%)] Loss: 0.000055 Train Epoch: 2 [41600/104000 (40%)] Loss: 0.000046 Train Epoch: 2 [42240/104000 (41%)] Loss: 0.000083 Train Epoch: 2 [42880/104000 (41%)] Loss: 0.000153 Train Epoch: 2 [43520/104000 (42%)] Loss: 0.000138 Train Epoch: 2 [44160/104000 (42%)] Loss: 0.000025 Train Epoch: 2 [44800/104000 (43%)] Loss: 0.000027 Train Epoch: 2 [45440/104000 (44%)] Loss: 0.000072 Train Epoch: 2 [46080/104000 (44%)] Loss: 0.000040 Train Epoch: 2 [46720/104000 (45%)] Loss: 0.000063 Train Epoch: 2 [47360/104000 (46%)] Loss: 0.000047 Train Epoch: 2 [48000/104000 (46%)] Loss: 0.000028 Train Epoch: 2 [48640/104000 (47%)] Loss: 0.000244 Train Epoch: 2 [49280/104000 (47%)] Loss: 0.000033 Train Epoch: 2 [49920/104000 (48%)] Loss: 0.000054 Train Epoch: 2 [50560/104000 (49%)] Loss: 0.000039 Train Epoch: 2 [51200/104000 (49%)] Loss: 0.000052 Train Epoch: 2 [51840/104000 (50%)] Loss: 0.000031 Train Epoch: 2 [52480/104000 (50%)] Loss: 0.000042 Train Epoch: 2 [53120/104000 (51%)] Loss: 0.000025 Train Epoch: 2 [53760/104000 (52%)] Loss: 0.000221 Train Epoch: 2 [54400/104000 (52%)] Loss: 0.000047 Train Epoch: 2 [55040/104000 (53%)] Loss: 0.000020 Train Epoch: 2 [55680/104000 (54%)] Loss: 0.000355 Train Epoch: 2 [56320/104000 (54%)] Loss: 0.000055 Train Epoch: 2 [56960/104000 (55%)] Loss: 0.000035 Train Epoch: 2 [57600/104000 (55%)] Loss: 0.000031 Train Epoch: 2 [58240/104000 (56%)] Loss: 0.000079 Train Epoch: 2 [58880/104000 (57%)] Loss: 0.000090 Train Epoch: 2 [59520/104000 (57%)] Loss: 0.000287 Train Epoch: 2 [60160/104000 (58%)] Loss: 0.000035 Train Epoch: 2 [60800/104000 (58%)] Loss: 0.000029 Train Epoch: 2 [61440/104000 (59%)] Loss: 0.000204 Train Epoch: 2 [62080/104000 (60%)] Loss: 0.000019 Train Epoch: 2 [62720/104000 (60%)] Loss: 0.000139 Train Epoch: 2 [63360/104000 (61%)] Loss: 0.000047 Train Epoch: 2 [64000/104000 (62%)] Loss: 0.000029 Train Epoch: 2 [64640/104000 (62%)] Loss: 0.000207 Train Epoch: 2 [65280/104000 (63%)] Loss: 0.000116 Train Epoch: 2 [65920/104000 (63%)] Loss: 0.000029 Train Epoch: 2 [66560/104000 (64%)] Loss: 0.000065 Train Epoch: 2 [67200/104000 (65%)] Loss: 0.000086 Train Epoch: 2 [67840/104000 (65%)] Loss: 0.000135 Train Epoch: 2 [68480/104000 (66%)] Loss: 0.000037 Train Epoch: 2 [69120/104000 (66%)] Loss: 0.000017 Train Epoch: 2 [69760/104000 (67%)] Loss: 0.000020 Train Epoch: 2 [70400/104000 (68%)] Loss: 0.000073 Train Epoch: 2 [71040/104000 (68%)] Loss: 0.000063 Train Epoch: 2 [71680/104000 (69%)] Loss: 0.000035 Train Epoch: 2 [72320/104000 (70%)] Loss: 0.000089 Train Epoch: 2 [72960/104000 (70%)] Loss: 0.000064 Train Epoch: 2 [73600/104000 (71%)] Loss: 0.000063 Train Epoch: 2 [74240/104000 (71%)] Loss: 0.000041 Train Epoch: 2 [74880/104000 (72%)] Loss: 0.000083 Train Epoch: 2 [75520/104000 (73%)] Loss: 0.000061 Train Epoch: 2 [76160/104000 (73%)] Loss: 0.000040 Train Epoch: 2 [76800/104000 (74%)] Loss: 0.000108 Train Epoch: 2 [77440/104000 (74%)] Loss: 0.000168 Train Epoch: 2 [78080/104000 (75%)] Loss: 0.000098 Train Epoch: 2 [78720/104000 (76%)] Loss: 0.000092 Train Epoch: 2 [79360/104000 (76%)] Loss: 0.000039 Train Epoch: 2 [80000/104000 (77%)] Loss: 0.000044 Train Epoch: 2 [80640/104000 (78%)] Loss: 0.000038 Train Epoch: 2 [81280/104000 (78%)] Loss: 0.000030 Train Epoch: 2 [81920/104000 (79%)] Loss: 0.000176 Train Epoch: 2 [82560/104000 (79%)] Loss: 0.000016 Train Epoch: 2 [83200/104000 (80%)] Loss: 0.000231 Train Epoch: 2 [83840/104000 (81%)] Loss: 0.000051 Train Epoch: 2 [84480/104000 (81%)] Loss: 0.000034 Train Epoch: 2 [85120/104000 (82%)] Loss: 0.000025 Train Epoch: 2 [85760/104000 (82%)] Loss: 0.000033 Train Epoch: 2 [86400/104000 (83%)] Loss: 0.000022 Train Epoch: 2 [87040/104000 (84%)] Loss: 0.000084 Train Epoch: 2 [87680/104000 (84%)] Loss: 0.000041 Train Epoch: 2 [88320/104000 (85%)] Loss: 0.000130 Train Epoch: 2 [88960/104000 (86%)] Loss: 0.000398 Train Epoch: 2 [89600/104000 (86%)] Loss: 0.000035 Train Epoch: 2 [90240/104000 (87%)] Loss: 0.000026 Train Epoch: 2 [90880/104000 (87%)] Loss: 0.000026 Train Epoch: 2 [91520/104000 (88%)] Loss: 0.000019 Train Epoch: 2 [92160/104000 (89%)] Loss: 0.000027 Train Epoch: 2 [92800/104000 (89%)] Loss: 0.000099 Train Epoch: 2 [93440/104000 (90%)] Loss: 0.000032 Train Epoch: 2 [94080/104000 (90%)] Loss: 0.000036 Train Epoch: 2 [94720/104000 (91%)] Loss: 0.000061 Train Epoch: 2 [95360/104000 (92%)] Loss: 0.000081 Train Epoch: 2 [96000/104000 (92%)] Loss: 0.000016 Train Epoch: 2 [96640/104000 (93%)] Loss: 0.000015 Train Epoch: 2 [97280/104000 (94%)] Loss: 0.000020 Train Epoch: 2 [97920/104000 (94%)] Loss: 0.000035 Train Epoch: 2 [98560/104000 (95%)] Loss: 0.000047 Train Epoch: 2 [99200/104000 (95%)] Loss: 0.000043 Train Epoch: 2 [99840/104000 (96%)] Loss: 0.000189 Train Epoch: 2 [100480/104000 (97%)] Loss: 0.000012 Train Epoch: 2 [101120/104000 (97%)] Loss: 0.000154 Train Epoch: 2 [101760/104000 (98%)] Loss: 0.000015 Train Epoch: 2 [102400/104000 (98%)] Loss: 0.000015 Train Epoch: 2 [103040/104000 (99%)] Loss: 0.000222 Train Epoch: 2 [103680/104000 (100%)] Loss: 0.000062 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2827, Accuracy: 25898/30000 (86%) Train Epoch: 3 [0/104000 (0%)] Loss: 0.000062 Train Epoch: 3 [640/104000 (1%)] Loss: 0.000032 Train Epoch: 3 [1280/104000 (1%)] Loss: 0.000027 Train Epoch: 3 [1920/104000 (2%)] Loss: 0.000102 Train Epoch: 3 [2560/104000 (2%)] Loss: 0.000049 Train Epoch: 3 [3200/104000 (3%)] Loss: 0.000057 Train Epoch: 3 [3840/104000 (4%)] Loss: 0.000086 Train Epoch: 3 [4480/104000 (4%)] Loss: 0.000031 Train Epoch: 3 [5120/104000 (5%)] Loss: 0.000087 Train Epoch: 3 [5760/104000 (6%)] Loss: 0.000037 Train Epoch: 3 [6400/104000 (6%)] Loss: 0.000068 Train Epoch: 3 [7040/104000 (7%)] Loss: 0.000019 Train Epoch: 3 [7680/104000 (7%)] Loss: 0.000057 Train Epoch: 3 [8320/104000 (8%)] Loss: 0.000036 Train Epoch: 3 [8960/104000 (9%)] Loss: 0.000053 Train Epoch: 3 [9600/104000 (9%)] Loss: 0.000274 Train Epoch: 3 [10240/104000 (10%)] Loss: 0.000014 Train Epoch: 3 [10880/104000 (10%)] Loss: 0.000030 Train Epoch: 3 [11520/104000 (11%)] Loss: 0.000088 Train Epoch: 3 [12160/104000 (12%)] Loss: 0.000022 Train Epoch: 3 [12800/104000 (12%)] Loss: 0.000152 Train Epoch: 3 [13440/104000 (13%)] Loss: 0.000109 Train Epoch: 3 [14080/104000 (14%)] Loss: 0.000042 Train Epoch: 3 [14720/104000 (14%)] Loss: 0.000021 Train Epoch: 3 [15360/104000 (15%)] Loss: 0.000037 Train Epoch: 3 [16000/104000 (15%)] Loss: 0.000025 Train Epoch: 3 [16640/104000 (16%)] Loss: 0.000032 Train Epoch: 3 [17280/104000 (17%)] Loss: 0.000028 Train Epoch: 3 [17920/104000 (17%)] Loss: 0.000023 Train Epoch: 3 [18560/104000 (18%)] Loss: 0.000031 Train Epoch: 3 [19200/104000 (18%)] Loss: 0.000041 Train Epoch: 3 [19840/104000 (19%)] Loss: 0.000016 Train Epoch: 3 [20480/104000 (20%)] Loss: 0.000025 Train Epoch: 3 [21120/104000 (20%)] Loss: 0.000094 Train Epoch: 3 [21760/104000 (21%)] Loss: 0.000069 Train Epoch: 3 [22400/104000 (22%)] Loss: 0.000120 Train Epoch: 3 [23040/104000 (22%)] Loss: 0.000043 Train Epoch: 3 [23680/104000 (23%)] Loss: 0.000028 Train Epoch: 3 [24320/104000 (23%)] Loss: 0.000026 Train Epoch: 3 [24960/104000 (24%)] Loss: 0.000016 Train Epoch: 3 [25600/104000 (25%)] Loss: 0.000033 Train Epoch: 3 [26240/104000 (25%)] Loss: 0.000031 Train Epoch: 3 [26880/104000 (26%)] Loss: 0.000011 Train Epoch: 3 [27520/104000 (26%)] Loss: 0.000067 Train Epoch: 3 [28160/104000 (27%)] Loss: 0.000037 Train Epoch: 3 [28800/104000 (28%)] Loss: 0.000104 Train Epoch: 3 [29440/104000 (28%)] Loss: 0.000057 Train Epoch: 3 [30080/104000 (29%)] Loss: 0.000020 Train Epoch: 3 [30720/104000 (30%)] Loss: 0.000024 Train Epoch: 3 [31360/104000 (30%)] Loss: 0.000063 Train Epoch: 3 [32000/104000 (31%)] Loss: 0.000027 Train Epoch: 3 [32640/104000 (31%)] Loss: 0.000036 Train Epoch: 3 [33280/104000 (32%)] Loss: 0.000017 Train Epoch: 3 [33920/104000 (33%)] Loss: 0.000062 Train Epoch: 3 [34560/104000 (33%)] Loss: 0.000018 Train Epoch: 3 [35200/104000 (34%)] Loss: 0.000043 Train Epoch: 3 [35840/104000 (34%)] Loss: 0.000016 Train Epoch: 3 [36480/104000 (35%)] Loss: 0.000019 Train Epoch: 3 [37120/104000 (36%)] Loss: 0.000100 Train Epoch: 3 [37760/104000 (36%)] Loss: 0.000013 Train Epoch: 3 [38400/104000 (37%)] Loss: 0.000170 Train Epoch: 3 [39040/104000 (38%)] Loss: 0.000072 Train Epoch: 3 [39680/104000 (38%)] Loss: 0.000037 Train Epoch: 3 [40320/104000 (39%)] Loss: 0.000023 Train Epoch: 3 [40960/104000 (39%)] Loss: 0.000012 Train Epoch: 3 [41600/104000 (40%)] Loss: 0.000027 Train Epoch: 3 [42240/104000 (41%)] Loss: 0.000034 Train Epoch: 3 [42880/104000 (41%)] Loss: 0.000018 Train Epoch: 3 [43520/104000 (42%)] Loss: 0.000017 Train Epoch: 3 [44160/104000 (42%)] Loss: 0.000143 Train Epoch: 3 [44800/104000 (43%)] Loss: 0.000163 Train Epoch: 3 [45440/104000 (44%)] Loss: 0.000014 Train Epoch: 3 [46080/104000 (44%)] Loss: 0.000017 Train Epoch: 3 [46720/104000 (45%)] Loss: 0.000014 Train Epoch: 3 [47360/104000 (46%)] Loss: 0.000045 Train Epoch: 3 [48000/104000 (46%)] Loss: 0.000134 Train Epoch: 3 [48640/104000 (47%)] Loss: 0.000019 Train Epoch: 3 [49280/104000 (47%)] Loss: 0.000125 Train Epoch: 3 [49920/104000 (48%)] Loss: 0.000016 Train Epoch: 3 [50560/104000 (49%)] Loss: 0.000064 Train Epoch: 3 [51200/104000 (49%)] Loss: 0.000034 Train Epoch: 3 [51840/104000 (50%)] Loss: 0.000051 Train Epoch: 3 [52480/104000 (50%)] Loss: 0.000149 Train Epoch: 3 [53120/104000 (51%)] Loss: 0.000018 Train Epoch: 3 [53760/104000 (52%)] Loss: 0.000008 Train Epoch: 3 [54400/104000 (52%)] Loss: 0.000020 Train Epoch: 3 [55040/104000 (53%)] Loss: 0.000005 Train Epoch: 3 [55680/104000 (54%)] Loss: 0.000012 Train Epoch: 3 [56320/104000 (54%)] Loss: 0.000017 Train Epoch: 3 [56960/104000 (55%)] Loss: 0.000024 Train Epoch: 3 [57600/104000 (55%)] Loss: 0.000100 Train Epoch: 3 [58240/104000 (56%)] Loss: 0.000027 Train Epoch: 3 [58880/104000 (57%)] Loss: 0.000125 Train Epoch: 3 [59520/104000 (57%)] Loss: 0.000013 Train Epoch: 3 [60160/104000 (58%)] Loss: 0.000011 Train Epoch: 3 [60800/104000 (58%)] Loss: 0.000019 Train Epoch: 3 [61440/104000 (59%)] Loss: 0.000019 Train Epoch: 3 [62080/104000 (60%)] Loss: 0.000015 Train Epoch: 3 [62720/104000 (60%)] Loss: 0.000013 Train Epoch: 3 [63360/104000 (61%)] Loss: 0.000028 Train Epoch: 3 [64000/104000 (62%)] Loss: 0.000008 Train Epoch: 3 [64640/104000 (62%)] Loss: 0.000013 Train Epoch: 3 [65280/104000 (63%)] Loss: 0.000024 Train Epoch: 3 [65920/104000 (63%)] Loss: 0.000040 Train Epoch: 3 [66560/104000 (64%)] Loss: 0.000021 Train Epoch: 3 [67200/104000 (65%)] Loss: 0.000017 Train Epoch: 3 [67840/104000 (65%)] Loss: 0.000027 Train Epoch: 3 [68480/104000 (66%)] Loss: 0.000102 Train Epoch: 3 [69120/104000 (66%)] Loss: 0.000017 Train Epoch: 3 [69760/104000 (67%)] Loss: 0.000042 Train Epoch: 3 [70400/104000 (68%)] Loss: 0.000033 Train Epoch: 3 [71040/104000 (68%)] Loss: 0.000131 Train Epoch: 3 [71680/104000 (69%)] Loss: 0.000275 Train Epoch: 3 [72320/104000 (70%)] Loss: 0.000014 Train Epoch: 3 [72960/104000 (70%)] Loss: 0.000013 Train Epoch: 3 [73600/104000 (71%)] Loss: 0.000013 Train Epoch: 3 [74240/104000 (71%)] Loss: 0.000365 Train Epoch: 3 [74880/104000 (72%)] Loss: 0.000345 Train Epoch: 3 [75520/104000 (73%)] Loss: 0.000024 Train Epoch: 3 [76160/104000 (73%)] Loss: 0.000019 Train Epoch: 3 [76800/104000 (74%)] Loss: 0.000037 Train Epoch: 3 [77440/104000 (74%)] Loss: 0.000045 Train Epoch: 3 [78080/104000 (75%)] Loss: 0.000022 Train Epoch: 3 [78720/104000 (76%)] Loss: 0.000131 Train Epoch: 3 [79360/104000 (76%)] Loss: 0.000036 Train Epoch: 3 [80000/104000 (77%)] Loss: 0.000440 Train Epoch: 3 [80640/104000 (78%)] Loss: 0.000014 Train Epoch: 3 [81280/104000 (78%)] Loss: 0.000013 Train Epoch: 3 [81920/104000 (79%)] Loss: 0.000020 Train Epoch: 3 [82560/104000 (79%)] Loss: 0.000081 Train Epoch: 3 [83200/104000 (80%)] Loss: 0.000016 Train Epoch: 3 [83840/104000 (81%)] Loss: 0.000031 Train Epoch: 3 [84480/104000 (81%)] Loss: 0.000012 Train Epoch: 3 [85120/104000 (82%)] Loss: 0.000022 Train Epoch: 3 [85760/104000 (82%)] Loss: 0.000040 Train Epoch: 3 [86400/104000 (83%)] Loss: 0.000138 Train Epoch: 3 [87040/104000 (84%)] Loss: 0.000037 Train Epoch: 3 [87680/104000 (84%)] Loss: 0.000013 Train Epoch: 3 [88320/104000 (85%)] Loss: 0.000011 Train Epoch: 3 [88960/104000 (86%)] Loss: 0.000017 Train Epoch: 3 [89600/104000 (86%)] Loss: 0.000011 Train Epoch: 3 [90240/104000 (87%)] Loss: 0.000108 Train Epoch: 3 [90880/104000 (87%)] Loss: 0.000011 Train Epoch: 3 [91520/104000 (88%)] Loss: 0.000022 Train Epoch: 3 [92160/104000 (89%)] Loss: 0.000009 Train Epoch: 3 [92800/104000 (89%)] Loss: 0.000128 Train Epoch: 3 [93440/104000 (90%)] Loss: 0.000032 Train Epoch: 3 [94080/104000 (90%)] Loss: 0.000076 Train Epoch: 3 [94720/104000 (91%)] Loss: 0.000048 Train Epoch: 3 [95360/104000 (92%)] Loss: 0.000016 Train Epoch: 3 [96000/104000 (92%)] Loss: 0.000017 Train Epoch: 3 [96640/104000 (93%)] Loss: 0.000045 Train Epoch: 3 [97280/104000 (94%)] Loss: 0.000012 Train Epoch: 3 [97920/104000 (94%)] Loss: 0.000064 Train Epoch: 3 [98560/104000 (95%)] Loss: 0.000007 Train Epoch: 3 [99200/104000 (95%)] Loss: 0.000150 Train Epoch: 3 [99840/104000 (96%)] Loss: 0.000035 Train Epoch: 3 [100480/104000 (97%)] Loss: 0.000023 Train Epoch: 3 [101120/104000 (97%)] Loss: 0.000030 Train Epoch: 3 [101760/104000 (98%)] Loss: 0.000011 Train Epoch: 3 [102400/104000 (98%)] Loss: 0.000008 Train Epoch: 3 [103040/104000 (99%)] Loss: 0.000104 Train Epoch: 3 [103680/104000 (100%)] Loss: 0.000076 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2302, Accuracy: 26714/30000 (89%) Train Epoch: 4 [0/104000 (0%)] Loss: 0.000044 Train Epoch: 4 [640/104000 (1%)] Loss: 0.000011 Train Epoch: 4 [1280/104000 (1%)] Loss: 0.000048 Train Epoch: 4 [1920/104000 (2%)] Loss: 0.000016 Train Epoch: 4 [2560/104000 (2%)] Loss: 0.000070 Train Epoch: 4 [3200/104000 (3%)] Loss: 0.000233 Train Epoch: 4 [3840/104000 (4%)] Loss: 0.000013 Train Epoch: 4 [4480/104000 (4%)] Loss: 0.000017 Train Epoch: 4 [5120/104000 (5%)] Loss: 0.000011 Train Epoch: 4 [5760/104000 (6%)] Loss: 0.000047 Train Epoch: 4 [6400/104000 (6%)] Loss: 0.000021 Train Epoch: 4 [7040/104000 (7%)] Loss: 0.000015 Train Epoch: 4 [7680/104000 (7%)] Loss: 0.000010 Train Epoch: 4 [8320/104000 (8%)] Loss: 0.000016 Train Epoch: 4 [8960/104000 (9%)] Loss: 0.000012 Train Epoch: 4 [9600/104000 (9%)] Loss: 0.000017 Train Epoch: 4 [10240/104000 (10%)] Loss: 0.000005 Train Epoch: 4 [10880/104000 (10%)] Loss: 0.000027 Train Epoch: 4 [11520/104000 (11%)] Loss: 0.000047 Train Epoch: 4 [12160/104000 (12%)] Loss: 0.000008 Train Epoch: 4 [12800/104000 (12%)] Loss: 0.000239 Train Epoch: 4 [13440/104000 (13%)] Loss: 0.000015 Train Epoch: 4 [14080/104000 (14%)] Loss: 0.000012 Train Epoch: 4 [14720/104000 (14%)] Loss: 0.000044 Train Epoch: 4 [15360/104000 (15%)] Loss: 0.000010 Train Epoch: 4 [16000/104000 (15%)] Loss: 0.000021 Train Epoch: 4 [16640/104000 (16%)] Loss: 0.000028 Train Epoch: 4 [17280/104000 (17%)] Loss: 0.000018 Train Epoch: 4 [17920/104000 (17%)] Loss: 0.000043 Train Epoch: 4 [18560/104000 (18%)] Loss: 0.000007 Train Epoch: 4 [19200/104000 (18%)] Loss: 0.000013 Train Epoch: 4 [19840/104000 (19%)] Loss: 0.000008 Train Epoch: 4 [20480/104000 (20%)] Loss: 0.000022 Train Epoch: 4 [21120/104000 (20%)] Loss: 0.000032 Train Epoch: 4 [21760/104000 (21%)] Loss: 0.000032 Train Epoch: 4 [22400/104000 (22%)] Loss: 0.000028 Train Epoch: 4 [23040/104000 (22%)] Loss: 0.000020 Train Epoch: 4 [23680/104000 (23%)] Loss: 0.000015 Train Epoch: 4 [24320/104000 (23%)] Loss: 0.000085 Train Epoch: 4 [24960/104000 (24%)] Loss: 0.000023 Train Epoch: 4 [25600/104000 (25%)] Loss: 0.000100 Train Epoch: 4 [26240/104000 (25%)] Loss: 0.000014 Train Epoch: 4 [26880/104000 (26%)] Loss: 0.000042 Train Epoch: 4 [27520/104000 (26%)] Loss: 0.000014 Train Epoch: 4 [28160/104000 (27%)] Loss: 0.001163 Train Epoch: 4 [28800/104000 (28%)] Loss: 0.000009 Train Epoch: 4 [29440/104000 (28%)] Loss: 0.000016 Train Epoch: 4 [30080/104000 (29%)] Loss: 0.000017 Train Epoch: 4 [30720/104000 (30%)] Loss: 0.000026 Train Epoch: 4 [31360/104000 (30%)] Loss: 0.000020 Train Epoch: 4 [32000/104000 (31%)] Loss: 0.000164 Train Epoch: 4 [32640/104000 (31%)] Loss: 0.000022 Train Epoch: 4 [33280/104000 (32%)] Loss: 0.000019 Train Epoch: 4 [33920/104000 (33%)] Loss: 0.000013 Train Epoch: 4 [34560/104000 (33%)] Loss: 0.000015 Train Epoch: 4 [35200/104000 (34%)] Loss: 0.000013 Train Epoch: 4 [35840/104000 (34%)] Loss: 0.000013 Train Epoch: 4 [36480/104000 (35%)] Loss: 0.000010 Train Epoch: 4 [37120/104000 (36%)] Loss: 0.000013 Train Epoch: 4 [37760/104000 (36%)] Loss: 0.000013 Train Epoch: 4 [38400/104000 (37%)] Loss: 0.000019 Train Epoch: 4 [39040/104000 (38%)] Loss: 0.000032 Train Epoch: 4 [39680/104000 (38%)] Loss: 0.000092 Train Epoch: 4 [40320/104000 (39%)] Loss: 0.000040 Train Epoch: 4 [40960/104000 (39%)] Loss: 0.000012 Train Epoch: 4 [41600/104000 (40%)] Loss: 0.000007 Train Epoch: 4 [42240/104000 (41%)] Loss: 0.000022 Train Epoch: 4 [42880/104000 (41%)] Loss: 0.000031 Train Epoch: 4 [43520/104000 (42%)] Loss: 0.000011 Train Epoch: 4 [44160/104000 (42%)] Loss: 0.000020 Train Epoch: 4 [44800/104000 (43%)] Loss: 0.000069 Train Epoch: 4 [45440/104000 (44%)] Loss: 0.000010 Train Epoch: 4 [46080/104000 (44%)] Loss: 0.000019 Train Epoch: 4 [46720/104000 (45%)] Loss: 0.000154 Train Epoch: 4 [47360/104000 (46%)] Loss: 0.000017 Train Epoch: 4 [48000/104000 (46%)] Loss: 0.000012 Train Epoch: 4 [48640/104000 (47%)] Loss: 0.000029 Train Epoch: 4 [49280/104000 (47%)] Loss: 0.000184 Train Epoch: 4 [49920/104000 (48%)] Loss: 0.000031 Train Epoch: 4 [50560/104000 (49%)] Loss: 0.000011 Train Epoch: 4 [51200/104000 (49%)] Loss: 0.000022 Train Epoch: 4 [51840/104000 (50%)] Loss: 0.000031 Train Epoch: 4 [52480/104000 (50%)] Loss: 0.000010 Train Epoch: 4 [53120/104000 (51%)] Loss: 0.000016 Train Epoch: 4 [53760/104000 (52%)] Loss: 0.000020 Train Epoch: 4 [54400/104000 (52%)] Loss: 0.000019 Train Epoch: 4 [55040/104000 (53%)] Loss: 0.000046 Train Epoch: 4 [55680/104000 (54%)] Loss: 0.000026 Train Epoch: 4 [56320/104000 (54%)] Loss: 0.000038 Train Epoch: 4 [56960/104000 (55%)] Loss: 0.000046 Train Epoch: 4 [57600/104000 (55%)] Loss: 0.000009 Train Epoch: 4 [58240/104000 (56%)] Loss: 0.000021 Train Epoch: 4 [58880/104000 (57%)] Loss: 0.000009 Train Epoch: 4 [59520/104000 (57%)] Loss: 0.000015 Train Epoch: 4 [60160/104000 (58%)] Loss: 0.000008 Train Epoch: 4 [60800/104000 (58%)] Loss: 0.000025 Train Epoch: 4 [61440/104000 (59%)] Loss: 0.000094 Train Epoch: 4 [62080/104000 (60%)] Loss: 0.000016 Train Epoch: 4 [62720/104000 (60%)] Loss: 0.000012 Train Epoch: 4 [63360/104000 (61%)] Loss: 0.000015 Train Epoch: 4 [64000/104000 (62%)] Loss: 0.000009 Train Epoch: 4 [64640/104000 (62%)] Loss: 0.000030 Train Epoch: 4 [65280/104000 (63%)] Loss: 0.000016 Train Epoch: 4 [65920/104000 (63%)] Loss: 0.000013 Train Epoch: 4 [66560/104000 (64%)] Loss: 0.000016 Train Epoch: 4 [67200/104000 (65%)] Loss: 0.000012 Train Epoch: 4 [67840/104000 (65%)] Loss: 0.000020 Train Epoch: 4 [68480/104000 (66%)] Loss: 0.000016 Train Epoch: 4 [69120/104000 (66%)] Loss: 0.000011 Train Epoch: 4 [69760/104000 (67%)] Loss: 0.000009 Train Epoch: 4 [70400/104000 (68%)] Loss: 0.000031 Train Epoch: 4 [71040/104000 (68%)] Loss: 0.000011 Train Epoch: 4 [71680/104000 (69%)] Loss: 0.000090 Train Epoch: 4 [72320/104000 (70%)] Loss: 0.000005 Train Epoch: 4 [72960/104000 (70%)] Loss: 0.000012 Train Epoch: 4 [73600/104000 (71%)] Loss: 0.000011 Train Epoch: 4 [74240/104000 (71%)] Loss: 0.000013 Train Epoch: 4 [74880/104000 (72%)] Loss: 0.000009 Train Epoch: 4 [75520/104000 (73%)] Loss: 0.000018 Train Epoch: 4 [76160/104000 (73%)] Loss: 0.000129 Train Epoch: 4 [76800/104000 (74%)] Loss: 0.000013 Train Epoch: 4 [77440/104000 (74%)] Loss: 0.000029 Train Epoch: 4 [78080/104000 (75%)] Loss: 0.000009 Train Epoch: 4 [78720/104000 (76%)] Loss: 0.000013 Train Epoch: 4 [79360/104000 (76%)] Loss: 0.000047 Train Epoch: 4 [80000/104000 (77%)] Loss: 0.000168 Train Epoch: 4 [80640/104000 (78%)] Loss: 0.000032 Train Epoch: 4 [81280/104000 (78%)] Loss: 0.000010 Train Epoch: 4 [81920/104000 (79%)] Loss: 0.000206 Train Epoch: 4 [82560/104000 (79%)] Loss: 0.000019 Train Epoch: 4 [83200/104000 (80%)] Loss: 0.000007 Train Epoch: 4 [83840/104000 (81%)] Loss: 0.000030 Train Epoch: 4 [84480/104000 (81%)] Loss: 0.000020 Train Epoch: 4 [85120/104000 (82%)] Loss: 0.000006 Train Epoch: 4 [85760/104000 (82%)] Loss: 0.000145 Train Epoch: 4 [86400/104000 (83%)] Loss: 0.000008 Train Epoch: 4 [87040/104000 (84%)] Loss: 0.000009 Train Epoch: 4 [87680/104000 (84%)] Loss: 0.000023 Train Epoch: 4 [88320/104000 (85%)] Loss: 0.000037 Train Epoch: 4 [88960/104000 (86%)] Loss: 0.000027 Train Epoch: 4 [89600/104000 (86%)] Loss: 0.000007 Train Epoch: 4 [90240/104000 (87%)] Loss: 0.000008 Train Epoch: 4 [90880/104000 (87%)] Loss: 0.000007 Train Epoch: 4 [91520/104000 (88%)] Loss: 0.000005 Train Epoch: 4 [92160/104000 (89%)] Loss: 0.000027 Train Epoch: 4 [92800/104000 (89%)] Loss: 0.000008 Train Epoch: 4 [93440/104000 (90%)] Loss: 0.000006 Train Epoch: 4 [94080/104000 (90%)] Loss: 0.000006 Train Epoch: 4 [94720/104000 (91%)] Loss: 0.000043 Train Epoch: 4 [95360/104000 (92%)] Loss: 0.000011 Train Epoch: 4 [96000/104000 (92%)] Loss: 0.000014 Train Epoch: 4 [96640/104000 (93%)] Loss: 0.000015 Train Epoch: 4 [97280/104000 (94%)] Loss: 0.000606 Train Epoch: 4 [97920/104000 (94%)] Loss: 0.000010 Train Epoch: 4 [98560/104000 (95%)] Loss: 0.000019 Train Epoch: 4 [99200/104000 (95%)] Loss: 0.000034 Train Epoch: 4 [99840/104000 (96%)] Loss: 0.000071 Train Epoch: 4 [100480/104000 (97%)] Loss: 0.000017 Train Epoch: 4 [101120/104000 (97%)] Loss: 0.000009 Train Epoch: 4 [101760/104000 (98%)] Loss: 0.000028 Train Epoch: 4 [102400/104000 (98%)] Loss: 0.000041 Train Epoch: 4 [103040/104000 (99%)] Loss: 0.000083 Train Epoch: 4 [103680/104000 (100%)] Loss: 0.000031 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2675, Accuracy: 26228/30000 (87%) Train Epoch: 5 [0/104000 (0%)] Loss: 0.000007 Train Epoch: 5 [640/104000 (1%)] Loss: 0.000041 Train Epoch: 5 [1280/104000 (1%)] Loss: 0.000008 Train Epoch: 5 [1920/104000 (2%)] Loss: 0.000036 Train Epoch: 5 [2560/104000 (2%)] Loss: 0.000021 Train Epoch: 5 [3200/104000 (3%)] Loss: 0.000008 Train Epoch: 5 [3840/104000 (4%)] Loss: 0.000028 Train Epoch: 5 [4480/104000 (4%)] Loss: 0.000006 Train Epoch: 5 [5120/104000 (5%)] Loss: 0.000010 Train Epoch: 5 [5760/104000 (6%)] Loss: 0.000012 Train Epoch: 5 [6400/104000 (6%)] Loss: 0.000018 Train Epoch: 5 [7040/104000 (7%)] Loss: 0.000009 Train Epoch: 5 [7680/104000 (7%)] Loss: 0.000039 Train Epoch: 5 [8320/104000 (8%)] Loss: 0.000024 Train Epoch: 5 [8960/104000 (9%)] Loss: 0.000007 Train Epoch: 5 [9600/104000 (9%)] Loss: 0.000015 Train Epoch: 5 [10240/104000 (10%)] Loss: 0.000010 Train Epoch: 5 [10880/104000 (10%)] Loss: 0.000049 Train Epoch: 5 [11520/104000 (11%)] Loss: 0.000212 Train Epoch: 5 [12160/104000 (12%)] Loss: 0.000006 Train Epoch: 5 [12800/104000 (12%)] Loss: 0.000050 Train Epoch: 5 [13440/104000 (13%)] Loss: 0.000028 Train Epoch: 5 [14080/104000 (14%)] Loss: 0.000006 Train Epoch: 5 [14720/104000 (14%)] Loss: 0.000069 Train Epoch: 5 [15360/104000 (15%)] Loss: 0.000053 Train Epoch: 5 [16000/104000 (15%)] Loss: 0.000056 Train Epoch: 5 [16640/104000 (16%)] Loss: 0.000007 Train Epoch: 5 [17280/104000 (17%)] Loss: 0.000014 Train Epoch: 5 [17920/104000 (17%)] Loss: 0.000007 Train Epoch: 5 [18560/104000 (18%)] Loss: 0.000039 Train Epoch: 5 [19200/104000 (18%)] Loss: 0.000129 Train Epoch: 5 [19840/104000 (19%)] Loss: 0.000129 Train Epoch: 5 [20480/104000 (20%)] Loss: 0.000007 Train Epoch: 5 [21120/104000 (20%)] Loss: 0.000004 Train Epoch: 5 [21760/104000 (21%)] Loss: 0.000152 Train Epoch: 5 [22400/104000 (22%)] Loss: 0.000016 Train Epoch: 5 [23040/104000 (22%)] Loss: 0.000135 Train Epoch: 5 [23680/104000 (23%)] Loss: 0.000028 Train Epoch: 5 [24320/104000 (23%)] Loss: 0.000009 Train Epoch: 5 [24960/104000 (24%)] Loss: 0.000038 Train Epoch: 5 [25600/104000 (25%)] Loss: 0.000011 Train Epoch: 5 [26240/104000 (25%)] Loss: 0.000037 Train Epoch: 5 [26880/104000 (26%)] Loss: 0.000010 Train Epoch: 5 [27520/104000 (26%)] Loss: 0.000148 Train Epoch: 5 [28160/104000 (27%)] Loss: 0.000010 Train Epoch: 5 [28800/104000 (28%)] Loss: 0.000007 Train Epoch: 5 [29440/104000 (28%)] Loss: 0.000074 Train Epoch: 5 [30080/104000 (29%)] Loss: 0.000012 Train Epoch: 5 [30720/104000 (30%)] Loss: 0.000011 Train Epoch: 5 [31360/104000 (30%)] Loss: 0.000004 Train Epoch: 5 [32000/104000 (31%)] Loss: 0.000009 Train Epoch: 5 [32640/104000 (31%)] Loss: 0.000005 Train Epoch: 5 [33280/104000 (32%)] Loss: 0.000016 Train Epoch: 5 [33920/104000 (33%)] Loss: 0.000018 Train Epoch: 5 [34560/104000 (33%)] Loss: 0.000005 Train Epoch: 5 [35200/104000 (34%)] Loss: 0.000099 Train Epoch: 5 [35840/104000 (34%)] Loss: 0.000007 Train Epoch: 5 [36480/104000 (35%)] Loss: 0.000010 Train Epoch: 5 [37120/104000 (36%)] Loss: 0.000011 Train Epoch: 5 [37760/104000 (36%)] Loss: 0.000010 Train Epoch: 5 [38400/104000 (37%)] Loss: 0.000005 Train Epoch: 5 [39040/104000 (38%)] Loss: 0.000011 Train Epoch: 5 [39680/104000 (38%)] Loss: 0.000092 Train Epoch: 5 [40320/104000 (39%)] Loss: 0.000021 Train Epoch: 5 [40960/104000 (39%)] Loss: 0.000015 Train Epoch: 5 [41600/104000 (40%)] Loss: 0.000048 Train Epoch: 5 [42240/104000 (41%)] Loss: 0.000017 Train Epoch: 5 [42880/104000 (41%)] Loss: 0.000011 Train Epoch: 5 [43520/104000 (42%)] Loss: 0.000023 Train Epoch: 5 [44160/104000 (42%)] Loss: 0.000005 Train Epoch: 5 [44800/104000 (43%)] Loss: 0.000006 Train Epoch: 5 [45440/104000 (44%)] Loss: 0.000006 Train Epoch: 5 [46080/104000 (44%)] Loss: 0.000045 Train Epoch: 5 [46720/104000 (45%)] Loss: 0.000008 Train Epoch: 5 [47360/104000 (46%)] Loss: 0.000102 Train Epoch: 5 [48000/104000 (46%)] Loss: 0.000014 Train Epoch: 5 [48640/104000 (47%)] Loss: 0.000264 Train Epoch: 5 [49280/104000 (47%)] Loss: 0.000121 Train Epoch: 5 [49920/104000 (48%)] Loss: 0.000008 Train Epoch: 5 [50560/104000 (49%)] Loss: 0.000014 Train Epoch: 5 [51200/104000 (49%)] Loss: 0.000006 Train Epoch: 5 [51840/104000 (50%)] Loss: 0.000008 Train Epoch: 5 [52480/104000 (50%)] Loss: 0.000027 Train Epoch: 5 [53120/104000 (51%)] Loss: 0.000009 Train Epoch: 5 [53760/104000 (52%)] Loss: 0.000010 Train Epoch: 5 [54400/104000 (52%)] Loss: 0.000007 Train Epoch: 5 [55040/104000 (53%)] Loss: 0.000025 Train Epoch: 5 [55680/104000 (54%)] Loss: 0.000319 Train Epoch: 5 [56320/104000 (54%)] Loss: 0.000014 Train Epoch: 5 [56960/104000 (55%)] Loss: 0.000090 Train Epoch: 5 [57600/104000 (55%)] Loss: 0.000015 Train Epoch: 5 [58240/104000 (56%)] Loss: 0.000012 Train Epoch: 5 [58880/104000 (57%)] Loss: 0.000005 Train Epoch: 5 [59520/104000 (57%)] Loss: 0.000005 Train Epoch: 5 [60160/104000 (58%)] Loss: 0.000009 Train Epoch: 5 [60800/104000 (58%)] Loss: 0.000004 Train Epoch: 5 [61440/104000 (59%)] Loss: 0.000007 Train Epoch: 5 [62080/104000 (60%)] Loss: 0.000011 Train Epoch: 5 [62720/104000 (60%)] Loss: 0.000056 Train Epoch: 5 [63360/104000 (61%)] Loss: 0.000004 Train Epoch: 5 [64000/104000 (62%)] Loss: 0.000020 Train Epoch: 5 [64640/104000 (62%)] Loss: 0.000007 Train Epoch: 5 [65280/104000 (63%)] Loss: 0.000007 Train Epoch: 5 [65920/104000 (63%)] Loss: 0.000006 Train Epoch: 5 [66560/104000 (64%)] Loss: 0.000005 Train Epoch: 5 [67200/104000 (65%)] Loss: 0.000010 Train Epoch: 5 [67840/104000 (65%)] Loss: 0.000008 Train Epoch: 5 [68480/104000 (66%)] Loss: 0.000003 Train Epoch: 5 [69120/104000 (66%)] Loss: 0.000007 Train Epoch: 5 [69760/104000 (67%)] Loss: 0.000008 Train Epoch: 5 [70400/104000 (68%)] Loss: 0.000041 Train Epoch: 5 [71040/104000 (68%)] Loss: 0.000101 Train Epoch: 5 [71680/104000 (69%)] Loss: 0.000037 Train Epoch: 5 [72320/104000 (70%)] Loss: 0.000152 Train Epoch: 5 [72960/104000 (70%)] Loss: 0.000005 Train Epoch: 5 [73600/104000 (71%)] Loss: 0.000027 Train Epoch: 5 [74240/104000 (71%)] Loss: 0.000075 Train Epoch: 5 [74880/104000 (72%)] Loss: 0.000018 Train Epoch: 5 [75520/104000 (73%)] Loss: 0.000016 Train Epoch: 5 [76160/104000 (73%)] Loss: 0.000011 Train Epoch: 5 [76800/104000 (74%)] Loss: 0.000030 Train Epoch: 5 [77440/104000 (74%)] Loss: 0.000010 Train Epoch: 5 [78080/104000 (75%)] Loss: 0.000016 Train Epoch: 5 [78720/104000 (76%)] Loss: 0.000020 Train Epoch: 5 [79360/104000 (76%)] Loss: 0.000009 Train Epoch: 5 [80000/104000 (77%)] Loss: 0.000004 Train Epoch: 5 [80640/104000 (78%)] Loss: 0.000022 Train Epoch: 5 [81280/104000 (78%)] Loss: 0.000017 Train Epoch: 5 [81920/104000 (79%)] Loss: 0.000007 Train Epoch: 5 [82560/104000 (79%)] Loss: 0.000021 Train Epoch: 5 [83200/104000 (80%)] Loss: 0.000077 Train Epoch: 5 [83840/104000 (81%)] Loss: 0.000032 Train Epoch: 5 [84480/104000 (81%)] Loss: 0.000082 Train Epoch: 5 [85120/104000 (82%)] Loss: 0.000011 Train Epoch: 5 [85760/104000 (82%)] Loss: 0.000005 Train Epoch: 5 [86400/104000 (83%)] Loss: 0.000006 Train Epoch: 5 [87040/104000 (84%)] Loss: 0.000044 Train Epoch: 5 [87680/104000 (84%)] Loss: 0.000006 Train Epoch: 5 [88320/104000 (85%)] Loss: 0.000008 Train Epoch: 5 [88960/104000 (86%)] Loss: 0.000003 Train Epoch: 5 [89600/104000 (86%)] Loss: 0.000012 Train Epoch: 5 [90240/104000 (87%)] Loss: 0.000006 Train Epoch: 5 [90880/104000 (87%)] Loss: 0.000006 Train Epoch: 5 [91520/104000 (88%)] Loss: 0.000016 Train Epoch: 5 [92160/104000 (89%)] Loss: 0.000060 Train Epoch: 5 [92800/104000 (89%)] Loss: 0.000008 Train Epoch: 5 [93440/104000 (90%)] Loss: 0.000006 Train Epoch: 5 [94080/104000 (90%)] Loss: 0.000015 Train Epoch: 5 [94720/104000 (91%)] Loss: 0.000014 Train Epoch: 5 [95360/104000 (92%)] Loss: 0.000018 Train Epoch: 5 [96000/104000 (92%)] Loss: 0.000006 Train Epoch: 5 [96640/104000 (93%)] Loss: 0.000009 Train Epoch: 5 [97280/104000 (94%)] Loss: 0.000008 Train Epoch: 5 [97920/104000 (94%)] Loss: 0.000025 Train Epoch: 5 [98560/104000 (95%)] Loss: 0.000006 Train Epoch: 5 [99200/104000 (95%)] Loss: 0.000008 Train Epoch: 5 [99840/104000 (96%)] Loss: 0.000038 Train Epoch: 5 [100480/104000 (97%)] Loss: 0.000004 Train Epoch: 5 [101120/104000 (97%)] Loss: 0.000020 Train Epoch: 5 [101760/104000 (98%)] Loss: 0.000006 Train Epoch: 5 [102400/104000 (98%)] Loss: 0.000007 Train Epoch: 5 [103040/104000 (99%)] Loss: 0.000044 Train Epoch: 5 [103680/104000 (100%)] Loss: 0.000092 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.1840, Accuracy: 27434/30000 (91%) [100.0, 100.0] [88.56, 91.44666666666667] The number of neurons in CNN layer is 10 Train Epoch: 1 [0/104000 (0%)] Loss: 0.552990 Train Epoch: 1 [640/104000 (1%)] Loss: 0.007941 Train Epoch: 1 [1280/104000 (1%)] Loss: 0.004193 Train Epoch: 1 [1920/104000 (2%)] Loss: 0.003162 Train Epoch: 1 [2560/104000 (2%)] Loss: 0.002266 Train Epoch: 1 [3200/104000 (3%)] Loss: 0.004642 Train Epoch: 1 [3840/104000 (4%)] Loss: 0.001600 Train Epoch: 1 [4480/104000 (4%)] Loss: 0.001422 Train Epoch: 1 [5120/104000 (5%)] Loss: 0.001106 Train Epoch: 1 [5760/104000 (6%)] Loss: 0.001111 Train Epoch: 1 [6400/104000 (6%)] Loss: 0.002817 Train Epoch: 1 [7040/104000 (7%)] Loss: 0.000887 Train Epoch: 1 [7680/104000 (7%)] Loss: 0.000959 Train Epoch: 1 [8320/104000 (8%)] Loss: 0.002973 Train Epoch: 1 [8960/104000 (9%)] Loss: 0.000815 Train Epoch: 1 [9600/104000 (9%)] Loss: 0.000668 Train Epoch: 1 [10240/104000 (10%)] Loss: 0.000793 Train Epoch: 1 [10880/104000 (10%)] Loss: 0.000626 Train Epoch: 1 [11520/104000 (11%)] Loss: 0.000544 Train Epoch: 1 [12160/104000 (12%)] Loss: 0.000493 Train Epoch: 1 [12800/104000 (12%)] Loss: 0.001317 Train Epoch: 1 [13440/104000 (13%)] Loss: 0.000536 Train Epoch: 1 [14080/104000 (14%)] Loss: 0.000696 Train Epoch: 1 [14720/104000 (14%)] Loss: 0.000400 Train Epoch: 1 [15360/104000 (15%)] Loss: 0.000392 Train Epoch: 1 [16000/104000 (15%)] Loss: 0.000243 Train Epoch: 1 [16640/104000 (16%)] Loss: 0.000226 Train Epoch: 1 [17280/104000 (17%)] Loss: 0.000268 Train Epoch: 1 [17920/104000 (17%)] Loss: 0.000282 Train Epoch: 1 [18560/104000 (18%)] Loss: 0.000365 Train Epoch: 1 [19200/104000 (18%)] Loss: 0.000205 Train Epoch: 1 [19840/104000 (19%)] Loss: 0.000307 Train Epoch: 1 [20480/104000 (20%)] Loss: 0.000428 Train Epoch: 1 [21120/104000 (20%)] Loss: 0.000367 Train Epoch: 1 [21760/104000 (21%)] Loss: 0.000241 Train Epoch: 1 [22400/104000 (22%)] Loss: 0.000208 Train Epoch: 1 [23040/104000 (22%)] Loss: 0.000198 Train Epoch: 1 [23680/104000 (23%)] Loss: 0.000603 Train Epoch: 1 [24320/104000 (23%)] Loss: 0.000469 Train Epoch: 1 [24960/104000 (24%)] Loss: 0.000194 Train Epoch: 1 [25600/104000 (25%)] Loss: 0.000243 Train Epoch: 1 [26240/104000 (25%)] Loss: 0.000303 Train Epoch: 1 [26880/104000 (26%)] Loss: 0.000310 Train Epoch: 1 [27520/104000 (26%)] Loss: 0.000347 Train Epoch: 1 [28160/104000 (27%)] Loss: 0.000274 Train Epoch: 1 [28800/104000 (28%)] Loss: 0.000292 Train Epoch: 1 [29440/104000 (28%)] Loss: 0.000304 Train Epoch: 1 [30080/104000 (29%)] Loss: 0.000176 Train Epoch: 1 [30720/104000 (30%)] Loss: 0.000490 Train Epoch: 1 [31360/104000 (30%)] Loss: 0.000162 Train Epoch: 1 [32000/104000 (31%)] Loss: 0.000216 Train Epoch: 1 [32640/104000 (31%)] Loss: 0.000513 Train Epoch: 1 [33280/104000 (32%)] Loss: 0.000212 Train Epoch: 1 [33920/104000 (33%)] Loss: 0.000553 Train Epoch: 1 [34560/104000 (33%)] Loss: 0.000168 Train Epoch: 1 [35200/104000 (34%)] Loss: 0.000135 Train Epoch: 1 [35840/104000 (34%)] Loss: 0.000179 Train Epoch: 1 [36480/104000 (35%)] Loss: 0.000207 Train Epoch: 1 [37120/104000 (36%)] Loss: 0.000681 Train Epoch: 1 [37760/104000 (36%)] Loss: 0.000148 Train Epoch: 1 [38400/104000 (37%)] Loss: 0.000171 Train Epoch: 1 [39040/104000 (38%)] Loss: 0.000230 Train Epoch: 1 [39680/104000 (38%)] Loss: 0.000408 Train Epoch: 1 [40320/104000 (39%)] Loss: 0.000163 Train Epoch: 1 [40960/104000 (39%)] Loss: 0.000141 Train Epoch: 1 [41600/104000 (40%)] Loss: 0.000120 Train Epoch: 1 [42240/104000 (41%)] Loss: 0.000175 Train Epoch: 1 [42880/104000 (41%)] Loss: 0.000258 Train Epoch: 1 [43520/104000 (42%)] Loss: 0.000279 Train Epoch: 1 [44160/104000 (42%)] Loss: 0.000507 Train Epoch: 1 [44800/104000 (43%)] Loss: 0.000255 Train Epoch: 1 [45440/104000 (44%)] Loss: 0.000092 Train Epoch: 1 [46080/104000 (44%)] Loss: 0.000383 Train Epoch: 1 [46720/104000 (45%)] Loss: 0.000096 Train Epoch: 1 [47360/104000 (46%)] Loss: 0.000065 Train Epoch: 1 [48000/104000 (46%)] Loss: 0.000147 Train Epoch: 1 [48640/104000 (47%)] Loss: 0.000246 Train Epoch: 1 [49280/104000 (47%)] Loss: 0.000117 Train Epoch: 1 [49920/104000 (48%)] Loss: 0.000432 Train Epoch: 1 [50560/104000 (49%)] Loss: 0.000127 Train Epoch: 1 [51200/104000 (49%)] Loss: 0.000390 Train Epoch: 1 [51840/104000 (50%)] Loss: 0.000071 Train Epoch: 1 [52480/104000 (50%)] Loss: 0.000879 Train Epoch: 1 [53120/104000 (51%)] Loss: 0.000102 Train Epoch: 1 [53760/104000 (52%)] Loss: 0.000092 Train Epoch: 1 [54400/104000 (52%)] Loss: 0.000086 Train Epoch: 1 [55040/104000 (53%)] Loss: 0.000228 Train Epoch: 1 [55680/104000 (54%)] Loss: 0.000131 Train Epoch: 1 [56320/104000 (54%)] Loss: 0.000146 Train Epoch: 1 [56960/104000 (55%)] Loss: 0.000064 Train Epoch: 1 [57600/104000 (55%)] Loss: 0.000046 Train Epoch: 1 [58240/104000 (56%)] Loss: 0.000090 Train Epoch: 1 [58880/104000 (57%)] Loss: 0.000709 Train Epoch: 1 [59520/104000 (57%)] Loss: 0.000289 Train Epoch: 1 [60160/104000 (58%)] Loss: 0.000095 Train Epoch: 1 [60800/104000 (58%)] Loss: 0.000108 Train Epoch: 1 [61440/104000 (59%)] Loss: 0.000055 Train Epoch: 1 [62080/104000 (60%)] Loss: 0.000140 Train Epoch: 1 [62720/104000 (60%)] Loss: 0.000154 Train Epoch: 1 [63360/104000 (61%)] Loss: 0.000128 Train Epoch: 1 [64000/104000 (62%)] Loss: 0.000132 Train Epoch: 1 [64640/104000 (62%)] Loss: 0.000061 Train Epoch: 1 [65280/104000 (63%)] Loss: 0.000073 Train Epoch: 1 [65920/104000 (63%)] Loss: 0.000036 Train Epoch: 1 [66560/104000 (64%)] Loss: 0.000077 Train Epoch: 1 [67200/104000 (65%)] Loss: 0.000076 Train Epoch: 1 [67840/104000 (65%)] Loss: 0.000640 Train Epoch: 1 [68480/104000 (66%)] Loss: 0.000084 Train Epoch: 1 [69120/104000 (66%)] Loss: 0.000177 Train Epoch: 1 [69760/104000 (67%)] Loss: 0.000173 Train Epoch: 1 [70400/104000 (68%)] Loss: 0.000070 Train Epoch: 1 [71040/104000 (68%)] Loss: 0.000653 Train Epoch: 1 [71680/104000 (69%)] Loss: 0.000108 Train Epoch: 1 [72320/104000 (70%)] Loss: 0.000086 Train Epoch: 1 [72960/104000 (70%)] Loss: 0.000351 Train Epoch: 1 [73600/104000 (71%)] Loss: 0.001071 Train Epoch: 1 [74240/104000 (71%)] Loss: 0.000356 Train Epoch: 1 [74880/104000 (72%)] Loss: 0.000096 Train Epoch: 1 [75520/104000 (73%)] Loss: 0.000082 Train Epoch: 1 [76160/104000 (73%)] Loss: 0.000059 Train Epoch: 1 [76800/104000 (74%)] Loss: 0.000364 Train Epoch: 1 [77440/104000 (74%)] Loss: 0.000427 Train Epoch: 1 [78080/104000 (75%)] Loss: 0.000066 Train Epoch: 1 [78720/104000 (76%)] Loss: 0.000175 Train Epoch: 1 [79360/104000 (76%)] Loss: 0.002394 Train Epoch: 1 [80000/104000 (77%)] Loss: 0.000264 Train Epoch: 1 [80640/104000 (78%)] Loss: 0.000103 Train Epoch: 1 [81280/104000 (78%)] Loss: 0.000052 Train Epoch: 1 [81920/104000 (79%)] Loss: 0.000079 Train Epoch: 1 [82560/104000 (79%)] Loss: 0.000042 Train Epoch: 1 [83200/104000 (80%)] Loss: 0.000111 Train Epoch: 1 [83840/104000 (81%)] Loss: 0.000067 Train Epoch: 1 [84480/104000 (81%)] Loss: 0.000086 Train Epoch: 1 [85120/104000 (82%)] Loss: 0.000658 Train Epoch: 1 [85760/104000 (82%)] Loss: 0.000236 Train Epoch: 1 [86400/104000 (83%)] Loss: 0.000054 Train Epoch: 1 [87040/104000 (84%)] Loss: 0.000112 Train Epoch: 1 [87680/104000 (84%)] Loss: 0.000034 Train Epoch: 1 [88320/104000 (85%)] Loss: 0.000068 Train Epoch: 1 [88960/104000 (86%)] Loss: 0.000254 Train Epoch: 1 [89600/104000 (86%)] Loss: 0.000058 Train Epoch: 1 [90240/104000 (87%)] Loss: 0.000086 Train Epoch: 1 [90880/104000 (87%)] Loss: 0.000078 Train Epoch: 1 [91520/104000 (88%)] Loss: 0.000045 Train Epoch: 1 [92160/104000 (89%)] Loss: 0.000104 Train Epoch: 1 [92800/104000 (89%)] Loss: 0.000059 Train Epoch: 1 [93440/104000 (90%)] Loss: 0.000071 Train Epoch: 1 [94080/104000 (90%)] Loss: 0.000035 Train Epoch: 1 [94720/104000 (91%)] Loss: 0.000094 Train Epoch: 1 [95360/104000 (92%)] Loss: 0.000089 Train Epoch: 1 [96000/104000 (92%)] Loss: 0.000071 Train Epoch: 1 [96640/104000 (93%)] Loss: 0.000202 Train Epoch: 1 [97280/104000 (94%)] Loss: 0.000061 Train Epoch: 1 [97920/104000 (94%)] Loss: 0.000074 Train Epoch: 1 [98560/104000 (95%)] Loss: 0.000062 Train Epoch: 1 [99200/104000 (95%)] Loss: 0.000194 Train Epoch: 1 [99840/104000 (96%)] Loss: 0.000070 Train Epoch: 1 [100480/104000 (97%)] Loss: 0.000037 Train Epoch: 1 [101120/104000 (97%)] Loss: 0.000226 Train Epoch: 1 [101760/104000 (98%)] Loss: 0.000044 Train Epoch: 1 [102400/104000 (98%)] Loss: 0.000114 Train Epoch: 1 [103040/104000 (99%)] Loss: 0.000205 Train Epoch: 1 [103680/104000 (100%)] Loss: 0.000196 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2509, Accuracy: 26222/30000 (87%) Train Epoch: 2 [0/104000 (0%)] Loss: 0.000063 Train Epoch: 2 [640/104000 (1%)] Loss: 0.000021 Train Epoch: 2 [1280/104000 (1%)] Loss: 0.000061 Train Epoch: 2 [1920/104000 (2%)] Loss: 0.000033 Train Epoch: 2 [2560/104000 (2%)] Loss: 0.000050 Train Epoch: 2 [3200/104000 (3%)] Loss: 0.000162 Train Epoch: 2 [3840/104000 (4%)] Loss: 0.000088 Train Epoch: 2 [4480/104000 (4%)] Loss: 0.000035 Train Epoch: 2 [5120/104000 (5%)] Loss: 0.000032 Train Epoch: 2 [5760/104000 (6%)] Loss: 0.000040 Train Epoch: 2 [6400/104000 (6%)] Loss: 0.000162 Train Epoch: 2 [7040/104000 (7%)] Loss: 0.000177 Train Epoch: 2 [7680/104000 (7%)] Loss: 0.000214 Train Epoch: 2 [8320/104000 (8%)] Loss: 0.000180 Train Epoch: 2 [8960/104000 (9%)] Loss: 0.000165 Train Epoch: 2 [9600/104000 (9%)] Loss: 0.000072 Train Epoch: 2 [10240/104000 (10%)] Loss: 0.000162 Train Epoch: 2 [10880/104000 (10%)] Loss: 0.000123 Train Epoch: 2 [11520/104000 (11%)] Loss: 0.001040 Train Epoch: 2 [12160/104000 (12%)] Loss: 0.000254 Train Epoch: 2 [12800/104000 (12%)] Loss: 0.000028 Train Epoch: 2 [13440/104000 (13%)] Loss: 0.000101 Train Epoch: 2 [14080/104000 (14%)] Loss: 0.000040 Train Epoch: 2 [14720/104000 (14%)] Loss: 0.000044 Train Epoch: 2 [15360/104000 (15%)] Loss: 0.000090 Train Epoch: 2 [16000/104000 (15%)] Loss: 0.000027 Train Epoch: 2 [16640/104000 (16%)] Loss: 0.000049 Train Epoch: 2 [17280/104000 (17%)] Loss: 0.000255 Train Epoch: 2 [17920/104000 (17%)] Loss: 0.000031 Train Epoch: 2 [18560/104000 (18%)] Loss: 0.000058 Train Epoch: 2 [19200/104000 (18%)] Loss: 0.000033 Train Epoch: 2 [19840/104000 (19%)] Loss: 0.000069 Train Epoch: 2 [20480/104000 (20%)] Loss: 0.000062 Train Epoch: 2 [21120/104000 (20%)] Loss: 0.000122 Train Epoch: 2 [21760/104000 (21%)] Loss: 0.000029 Train Epoch: 2 [22400/104000 (22%)] Loss: 0.000041 Train Epoch: 2 [23040/104000 (22%)] Loss: 0.000041 Train Epoch: 2 [23680/104000 (23%)] Loss: 0.000193 Train Epoch: 2 [24320/104000 (23%)] Loss: 0.000052 Train Epoch: 2 [24960/104000 (24%)] Loss: 0.000081 Train Epoch: 2 [25600/104000 (25%)] Loss: 0.000166 Train Epoch: 2 [26240/104000 (25%)] Loss: 0.000352 Train Epoch: 2 [26880/104000 (26%)] Loss: 0.000046 Train Epoch: 2 [27520/104000 (26%)] Loss: 0.000048 Train Epoch: 2 [28160/104000 (27%)] Loss: 0.000032 Train Epoch: 2 [28800/104000 (28%)] Loss: 0.000053 Train Epoch: 2 [29440/104000 (28%)] Loss: 0.000113 Train Epoch: 2 [30080/104000 (29%)] Loss: 0.000038 Train Epoch: 2 [30720/104000 (30%)] Loss: 0.000189 Train Epoch: 2 [31360/104000 (30%)] Loss: 0.000024 Train Epoch: 2 [32000/104000 (31%)] Loss: 0.000050 Train Epoch: 2 [32640/104000 (31%)] Loss: 0.000031 Train Epoch: 2 [33280/104000 (32%)] Loss: 0.000033 Train Epoch: 2 [33920/104000 (33%)] Loss: 0.000019 Train Epoch: 2 [34560/104000 (33%)] Loss: 0.000031 Train Epoch: 2 [35200/104000 (34%)] Loss: 0.000018 Train Epoch: 2 [35840/104000 (34%)] Loss: 0.000035 Train Epoch: 2 [36480/104000 (35%)] Loss: 0.000073 Train Epoch: 2 [37120/104000 (36%)] Loss: 0.000027 Train Epoch: 2 [37760/104000 (36%)] Loss: 0.000078 Train Epoch: 2 [38400/104000 (37%)] Loss: 0.000035 Train Epoch: 2 [39040/104000 (38%)] Loss: 0.000155 Train Epoch: 2 [39680/104000 (38%)] Loss: 0.000087 Train Epoch: 2 [40320/104000 (39%)] Loss: 0.000079 Train Epoch: 2 [40960/104000 (39%)] Loss: 0.000032 Train Epoch: 2 [41600/104000 (40%)] Loss: 0.000029 Train Epoch: 2 [42240/104000 (41%)] Loss: 0.000381 Train Epoch: 2 [42880/104000 (41%)] Loss: 0.000389 Train Epoch: 2 [43520/104000 (42%)] Loss: 0.000030 Train Epoch: 2 [44160/104000 (42%)] Loss: 0.000078 Train Epoch: 2 [44800/104000 (43%)] Loss: 0.000212 Train Epoch: 2 [45440/104000 (44%)] Loss: 0.000054 Train Epoch: 2 [46080/104000 (44%)] Loss: 0.000039 Train Epoch: 2 [46720/104000 (45%)] Loss: 0.000423 Train Epoch: 2 [47360/104000 (46%)] Loss: 0.000092 Train Epoch: 2 [48000/104000 (46%)] Loss: 0.000078 Train Epoch: 2 [48640/104000 (47%)] Loss: 0.000043 Train Epoch: 2 [49280/104000 (47%)] Loss: 0.000024 Train Epoch: 2 [49920/104000 (48%)] Loss: 0.000043 Train Epoch: 2 [50560/104000 (49%)] Loss: 0.000034 Train Epoch: 2 [51200/104000 (49%)] Loss: 0.000050 Train Epoch: 2 [51840/104000 (50%)] Loss: 0.000021 Train Epoch: 2 [52480/104000 (50%)] Loss: 0.000034 Train Epoch: 2 [53120/104000 (51%)] Loss: 0.000017 Train Epoch: 2 [53760/104000 (52%)] Loss: 0.000113 Train Epoch: 2 [54400/104000 (52%)] Loss: 0.000026 Train Epoch: 2 [55040/104000 (53%)] Loss: 0.000175 Train Epoch: 2 [55680/104000 (54%)] Loss: 0.000030 Train Epoch: 2 [56320/104000 (54%)] Loss: 0.000022 Train Epoch: 2 [56960/104000 (55%)] Loss: 0.000033 Train Epoch: 2 [57600/104000 (55%)] Loss: 0.000039 Train Epoch: 2 [58240/104000 (56%)] Loss: 0.000300 Train Epoch: 2 [58880/104000 (57%)] Loss: 0.000085 Train Epoch: 2 [59520/104000 (57%)] Loss: 0.000020 Train Epoch: 2 [60160/104000 (58%)] Loss: 0.000019 Train Epoch: 2 [60800/104000 (58%)] Loss: 0.000030 Train Epoch: 2 [61440/104000 (59%)] Loss: 0.000020 Train Epoch: 2 [62080/104000 (60%)] Loss: 0.000033 Train Epoch: 2 [62720/104000 (60%)] Loss: 0.000115 Train Epoch: 2 [63360/104000 (61%)] Loss: 0.000040 Train Epoch: 2 [64000/104000 (62%)] Loss: 0.000037 Train Epoch: 2 [64640/104000 (62%)] Loss: 0.000128 Train Epoch: 2 [65280/104000 (63%)] Loss: 0.000017 Train Epoch: 2 [65920/104000 (63%)] Loss: 0.000029 Train Epoch: 2 [66560/104000 (64%)] Loss: 0.000038 Train Epoch: 2 [67200/104000 (65%)] Loss: 0.000085 Train Epoch: 2 [67840/104000 (65%)] Loss: 0.000018 Train Epoch: 2 [68480/104000 (66%)] Loss: 0.000035 Train Epoch: 2 [69120/104000 (66%)] Loss: 0.000018 Train Epoch: 2 [69760/104000 (67%)] Loss: 0.000039 Train Epoch: 2 [70400/104000 (68%)] Loss: 0.000032 Train Epoch: 2 [71040/104000 (68%)] Loss: 0.000022 Train Epoch: 2 [71680/104000 (69%)] Loss: 0.000065 Train Epoch: 2 [72320/104000 (70%)] Loss: 0.000018 Train Epoch: 2 [72960/104000 (70%)] Loss: 0.000109 Train Epoch: 2 [73600/104000 (71%)] Loss: 0.000032 Train Epoch: 2 [74240/104000 (71%)] Loss: 0.000052 Train Epoch: 2 [74880/104000 (72%)] Loss: 0.000016 Train Epoch: 2 [75520/104000 (73%)] Loss: 0.000040 Train Epoch: 2 [76160/104000 (73%)] Loss: 0.000012 Train Epoch: 2 [76800/104000 (74%)] Loss: 0.000019 Train Epoch: 2 [77440/104000 (74%)] Loss: 0.000031 Train Epoch: 2 [78080/104000 (75%)] Loss: 0.000105 Train Epoch: 2 [78720/104000 (76%)] Loss: 0.000023 Train Epoch: 2 [79360/104000 (76%)] Loss: 0.000011 Train Epoch: 2 [80000/104000 (77%)] Loss: 0.000022 Train Epoch: 2 [80640/104000 (78%)] Loss: 0.000539 Train Epoch: 2 [81280/104000 (78%)] Loss: 0.000010 Train Epoch: 2 [81920/104000 (79%)] Loss: 0.000049 Train Epoch: 2 [82560/104000 (79%)] Loss: 0.000022 Train Epoch: 2 [83200/104000 (80%)] Loss: 0.000019 Train Epoch: 2 [83840/104000 (81%)] Loss: 0.000025 Train Epoch: 2 [84480/104000 (81%)] Loss: 0.000018 Train Epoch: 2 [85120/104000 (82%)] Loss: 0.000029 Train Epoch: 2 [85760/104000 (82%)] Loss: 0.000024 Train Epoch: 2 [86400/104000 (83%)] Loss: 0.000064 Train Epoch: 2 [87040/104000 (84%)] Loss: 0.000017 Train Epoch: 2 [87680/104000 (84%)] Loss: 0.000039 Train Epoch: 2 [88320/104000 (85%)] Loss: 0.000042 Train Epoch: 2 [88960/104000 (86%)] Loss: 0.000021 Train Epoch: 2 [89600/104000 (86%)] Loss: 0.000062 Train Epoch: 2 [90240/104000 (87%)] Loss: 0.000053 Train Epoch: 2 [90880/104000 (87%)] Loss: 0.000026 Train Epoch: 2 [91520/104000 (88%)] Loss: 0.000050 Train Epoch: 2 [92160/104000 (89%)] Loss: 0.000041 Train Epoch: 2 [92800/104000 (89%)] Loss: 0.000024 Train Epoch: 2 [93440/104000 (90%)] Loss: 0.000024 Train Epoch: 2 [94080/104000 (90%)] Loss: 0.000011 Train Epoch: 2 [94720/104000 (91%)] Loss: 0.000027 Train Epoch: 2 [95360/104000 (92%)] Loss: 0.000136 Train Epoch: 2 [96000/104000 (92%)] Loss: 0.000015 Train Epoch: 2 [96640/104000 (93%)] Loss: 0.000066 Train Epoch: 2 [97280/104000 (94%)] Loss: 0.000221 Train Epoch: 2 [97920/104000 (94%)] Loss: 0.000067 Train Epoch: 2 [98560/104000 (95%)] Loss: 0.000014 Train Epoch: 2 [99200/104000 (95%)] Loss: 0.000091 Train Epoch: 2 [99840/104000 (96%)] Loss: 0.000065 Train Epoch: 2 [100480/104000 (97%)] Loss: 0.000011 Train Epoch: 2 [101120/104000 (97%)] Loss: 0.000034 Train Epoch: 2 [101760/104000 (98%)] Loss: 0.000043 Train Epoch: 2 [102400/104000 (98%)] Loss: 0.000216 Train Epoch: 2 [103040/104000 (99%)] Loss: 0.000035 Train Epoch: 2 [103680/104000 (100%)] Loss: 0.000032 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2632, Accuracy: 26118/30000 (87%) Train Epoch: 3 [0/104000 (0%)] Loss: 0.000013 Train Epoch: 3 [640/104000 (1%)] Loss: 0.000040 Train Epoch: 3 [1280/104000 (1%)] Loss: 0.000020 Train Epoch: 3 [1920/104000 (2%)] Loss: 0.000028 Train Epoch: 3 [2560/104000 (2%)] Loss: 0.000020 Train Epoch: 3 [3200/104000 (3%)] Loss: 0.000021 Train Epoch: 3 [3840/104000 (4%)] Loss: 0.000022 Train Epoch: 3 [4480/104000 (4%)] Loss: 0.000078 Train Epoch: 3 [5120/104000 (5%)] Loss: 0.000067 Train Epoch: 3 [5760/104000 (6%)] Loss: 0.000024 Train Epoch: 3 [6400/104000 (6%)] Loss: 0.000128 Train Epoch: 3 [7040/104000 (7%)] Loss: 0.000016 Train Epoch: 3 [7680/104000 (7%)] Loss: 0.000036 Train Epoch: 3 [8320/104000 (8%)] Loss: 0.000023 Train Epoch: 3 [8960/104000 (9%)] Loss: 0.000034 Train Epoch: 3 [9600/104000 (9%)] Loss: 0.000033 Train Epoch: 3 [10240/104000 (10%)] Loss: 0.000048 Train Epoch: 3 [10880/104000 (10%)] Loss: 0.000074 Train Epoch: 3 [11520/104000 (11%)] Loss: 0.000023 Train Epoch: 3 [12160/104000 (12%)] Loss: 0.000009 Train Epoch: 3 [12800/104000 (12%)] Loss: 0.000012 Train Epoch: 3 [13440/104000 (13%)] Loss: 0.000043 Train Epoch: 3 [14080/104000 (14%)] Loss: 0.000050 Train Epoch: 3 [14720/104000 (14%)] Loss: 0.000025 Train Epoch: 3 [15360/104000 (15%)] Loss: 0.000066 Train Epoch: 3 [16000/104000 (15%)] Loss: 0.000021 Train Epoch: 3 [16640/104000 (16%)] Loss: 0.000028 Train Epoch: 3 [17280/104000 (17%)] Loss: 0.000069 Train Epoch: 3 [17920/104000 (17%)] Loss: 0.000577 Train Epoch: 3 [18560/104000 (18%)] Loss: 0.000056 Train Epoch: 3 [19200/104000 (18%)] Loss: 0.000051 Train Epoch: 3 [19840/104000 (19%)] Loss: 0.000024 Train Epoch: 3 [20480/104000 (20%)] Loss: 0.000044 Train Epoch: 3 [21120/104000 (20%)] Loss: 0.000008 Train Epoch: 3 [21760/104000 (21%)] Loss: 0.000084 Train Epoch: 3 [22400/104000 (22%)] Loss: 0.000015 Train Epoch: 3 [23040/104000 (22%)] Loss: 0.000048 Train Epoch: 3 [23680/104000 (23%)] Loss: 0.000087 Train Epoch: 3 [24320/104000 (23%)] Loss: 0.000081 Train Epoch: 3 [24960/104000 (24%)] Loss: 0.000085 Train Epoch: 3 [25600/104000 (25%)] Loss: 0.000049 Train Epoch: 3 [26240/104000 (25%)] Loss: 0.000016 Train Epoch: 3 [26880/104000 (26%)] Loss: 0.000944 Train Epoch: 3 [27520/104000 (26%)] Loss: 0.000015 Train Epoch: 3 [28160/104000 (27%)] Loss: 0.000020 Train Epoch: 3 [28800/104000 (28%)] Loss: 0.000038 Train Epoch: 3 [29440/104000 (28%)] Loss: 0.000048 Train Epoch: 3 [30080/104000 (29%)] Loss: 0.000017 Train Epoch: 3 [30720/104000 (30%)] Loss: 0.000014 Train Epoch: 3 [31360/104000 (30%)] Loss: 0.000011 Train Epoch: 3 [32000/104000 (31%)] Loss: 0.000208 Train Epoch: 3 [32640/104000 (31%)] Loss: 0.000019 Train Epoch: 3 [33280/104000 (32%)] Loss: 0.000092 Train Epoch: 3 [33920/104000 (33%)] Loss: 0.000065 Train Epoch: 3 [34560/104000 (33%)] Loss: 0.000018 Train Epoch: 3 [35200/104000 (34%)] Loss: 0.000012 Train Epoch: 3 [35840/104000 (34%)] Loss: 0.000012 Train Epoch: 3 [36480/104000 (35%)] Loss: 0.000495 Train Epoch: 3 [37120/104000 (36%)] Loss: 0.000007 Train Epoch: 3 [37760/104000 (36%)] Loss: 0.000019 Train Epoch: 3 [38400/104000 (37%)] Loss: 0.000009 Train Epoch: 3 [39040/104000 (38%)] Loss: 0.000019 Train Epoch: 3 [39680/104000 (38%)] Loss: 0.000014 Train Epoch: 3 [40320/104000 (39%)] Loss: 0.000032 Train Epoch: 3 [40960/104000 (39%)] Loss: 0.000062 Train Epoch: 3 [41600/104000 (40%)] Loss: 0.000005 Train Epoch: 3 [42240/104000 (41%)] Loss: 0.000022 Train Epoch: 3 [42880/104000 (41%)] Loss: 0.000134 Train Epoch: 3 [43520/104000 (42%)] Loss: 0.000119 Train Epoch: 3 [44160/104000 (42%)] Loss: 0.000040 Train Epoch: 3 [44800/104000 (43%)] Loss: 0.000053 Train Epoch: 3 [45440/104000 (44%)] Loss: 0.000101 Train Epoch: 3 [46080/104000 (44%)] Loss: 0.000548 Train Epoch: 3 [46720/104000 (45%)] Loss: 0.000008 Train Epoch: 3 [47360/104000 (46%)] Loss: 0.000022 Train Epoch: 3 [48000/104000 (46%)] Loss: 0.000016 Train Epoch: 3 [48640/104000 (47%)] Loss: 0.000016 Train Epoch: 3 [49280/104000 (47%)] Loss: 0.000215 Train Epoch: 3 [49920/104000 (48%)] Loss: 0.000015 Train Epoch: 3 [50560/104000 (49%)] Loss: 0.000349 Train Epoch: 3 [51200/104000 (49%)] Loss: 0.000023 Train Epoch: 3 [51840/104000 (50%)] Loss: 0.000019 Train Epoch: 3 [52480/104000 (50%)] Loss: 0.000029 Train Epoch: 3 [53120/104000 (51%)] Loss: 0.000146 Train Epoch: 3 [53760/104000 (52%)] Loss: 0.000016 Train Epoch: 3 [54400/104000 (52%)] Loss: 0.000019 Train Epoch: 3 [55040/104000 (53%)] Loss: 0.000020 Train Epoch: 3 [55680/104000 (54%)] Loss: 0.000048 Train Epoch: 3 [56320/104000 (54%)] Loss: 0.000187 Train Epoch: 3 [56960/104000 (55%)] Loss: 0.000015 Train Epoch: 3 [57600/104000 (55%)] Loss: 0.000008 Train Epoch: 3 [58240/104000 (56%)] Loss: 0.000007 Train Epoch: 3 [58880/104000 (57%)] Loss: 0.000012 Train Epoch: 3 [59520/104000 (57%)] Loss: 0.000012 Train Epoch: 3 [60160/104000 (58%)] Loss: 0.000100 Train Epoch: 3 [60800/104000 (58%)] Loss: 0.000092 Train Epoch: 3 [61440/104000 (59%)] Loss: 0.000009 Train Epoch: 3 [62080/104000 (60%)] Loss: 0.000222 Train Epoch: 3 [62720/104000 (60%)] Loss: 0.000231 Train Epoch: 3 [63360/104000 (61%)] Loss: 0.000017 Train Epoch: 3 [64000/104000 (62%)] Loss: 0.000010 Train Epoch: 3 [64640/104000 (62%)] Loss: 0.000045 Train Epoch: 3 [65280/104000 (63%)] Loss: 0.000136 Train Epoch: 3 [65920/104000 (63%)] Loss: 0.000020 Train Epoch: 3 [66560/104000 (64%)] Loss: 0.000018 Train Epoch: 3 [67200/104000 (65%)] Loss: 0.000019 Train Epoch: 3 [67840/104000 (65%)] Loss: 0.000028 Train Epoch: 3 [68480/104000 (66%)] Loss: 0.000013 Train Epoch: 3 [69120/104000 (66%)] Loss: 0.000025 Train Epoch: 3 [69760/104000 (67%)] Loss: 0.000026 Train Epoch: 3 [70400/104000 (68%)] Loss: 0.000013 Train Epoch: 3 [71040/104000 (68%)] Loss: 0.000233 Train Epoch: 3 [71680/104000 (69%)] Loss: 0.000016 Train Epoch: 3 [72320/104000 (70%)] Loss: 0.000092 Train Epoch: 3 [72960/104000 (70%)] Loss: 0.000115 Train Epoch: 3 [73600/104000 (71%)] Loss: 0.000017 Train Epoch: 3 [74240/104000 (71%)] Loss: 0.000011 Train Epoch: 3 [74880/104000 (72%)] Loss: 0.000066 Train Epoch: 3 [75520/104000 (73%)] Loss: 0.000013 Train Epoch: 3 [76160/104000 (73%)] Loss: 0.000048 Train Epoch: 3 [76800/104000 (74%)] Loss: 0.000009 Train Epoch: 3 [77440/104000 (74%)] Loss: 0.000016 Train Epoch: 3 [78080/104000 (75%)] Loss: 0.000031 Train Epoch: 3 [78720/104000 (76%)] Loss: 0.000025 Train Epoch: 3 [79360/104000 (76%)] Loss: 0.000018 Train Epoch: 3 [80000/104000 (77%)] Loss: 0.000115 Train Epoch: 3 [80640/104000 (78%)] Loss: 0.000021 Train Epoch: 3 [81280/104000 (78%)] Loss: 0.000046 Train Epoch: 3 [81920/104000 (79%)] Loss: 0.000012 Train Epoch: 3 [82560/104000 (79%)] Loss: 0.000011 Train Epoch: 3 [83200/104000 (80%)] Loss: 0.000027 Train Epoch: 3 [83840/104000 (81%)] Loss: 0.000042 Train Epoch: 3 [84480/104000 (81%)] Loss: 0.000056 Train Epoch: 3 [85120/104000 (82%)] Loss: 0.000012 Train Epoch: 3 [85760/104000 (82%)] Loss: 0.000045 Train Epoch: 3 [86400/104000 (83%)] Loss: 0.000017 Train Epoch: 3 [87040/104000 (84%)] Loss: 0.000027 Train Epoch: 3 [87680/104000 (84%)] Loss: 0.000012 Train Epoch: 3 [88320/104000 (85%)] Loss: 0.000058 Train Epoch: 3 [88960/104000 (86%)] Loss: 0.000018 Train Epoch: 3 [89600/104000 (86%)] Loss: 0.000019 Train Epoch: 3 [90240/104000 (87%)] Loss: 0.000014 Train Epoch: 3 [90880/104000 (87%)] Loss: 0.000009 Train Epoch: 3 [91520/104000 (88%)] Loss: 0.000008 Train Epoch: 3 [92160/104000 (89%)] Loss: 0.000008 Train Epoch: 3 [92800/104000 (89%)] Loss: 0.000094 Train Epoch: 3 [93440/104000 (90%)] Loss: 0.000103 Train Epoch: 3 [94080/104000 (90%)] Loss: 0.000040 Train Epoch: 3 [94720/104000 (91%)] Loss: 0.000068 Train Epoch: 3 [95360/104000 (92%)] Loss: 0.000033 Train Epoch: 3 [96000/104000 (92%)] Loss: 0.000008 Train Epoch: 3 [96640/104000 (93%)] Loss: 0.000018 Train Epoch: 3 [97280/104000 (94%)] Loss: 0.000122 Train Epoch: 3 [97920/104000 (94%)] Loss: 0.000055 Train Epoch: 3 [98560/104000 (95%)] Loss: 0.000018 Train Epoch: 3 [99200/104000 (95%)] Loss: 0.000033 Train Epoch: 3 [99840/104000 (96%)] Loss: 0.000014 Train Epoch: 3 [100480/104000 (97%)] Loss: 0.000009 Train Epoch: 3 [101120/104000 (97%)] Loss: 0.000012 Train Epoch: 3 [101760/104000 (98%)] Loss: 0.000015 Train Epoch: 3 [102400/104000 (98%)] Loss: 0.000009 Train Epoch: 3 [103040/104000 (99%)] Loss: 0.000008 Train Epoch: 3 [103680/104000 (100%)] Loss: 0.000037 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.1820, Accuracy: 27467/30000 (92%) Train Epoch: 4 [0/104000 (0%)] Loss: 0.000022 Train Epoch: 4 [640/104000 (1%)] Loss: 0.000015 Train Epoch: 4 [1280/104000 (1%)] Loss: 0.000029 Train Epoch: 4 [1920/104000 (2%)] Loss: 0.000010 Train Epoch: 4 [2560/104000 (2%)] Loss: 0.000053 Train Epoch: 4 [3200/104000 (3%)] Loss: 0.000009 Train Epoch: 4 [3840/104000 (4%)] Loss: 0.000012 Train Epoch: 4 [4480/104000 (4%)] Loss: 0.000034 Train Epoch: 4 [5120/104000 (5%)] Loss: 0.000009 Train Epoch: 4 [5760/104000 (6%)] Loss: 0.000009 Train Epoch: 4 [6400/104000 (6%)] Loss: 0.000095 Train Epoch: 4 [7040/104000 (7%)] Loss: 0.000007 Train Epoch: 4 [7680/104000 (7%)] Loss: 0.000044 Train Epoch: 4 [8320/104000 (8%)] Loss: 0.000015 Train Epoch: 4 [8960/104000 (9%)] Loss: 0.000021 Train Epoch: 4 [9600/104000 (9%)] Loss: 0.000014 Train Epoch: 4 [10240/104000 (10%)] Loss: 0.000030 Train Epoch: 4 [10880/104000 (10%)] Loss: 0.000013 Train Epoch: 4 [11520/104000 (11%)] Loss: 0.000021 Train Epoch: 4 [12160/104000 (12%)] Loss: 0.000007 Train Epoch: 4 [12800/104000 (12%)] Loss: 0.000043 Train Epoch: 4 [13440/104000 (13%)] Loss: 0.000027 Train Epoch: 4 [14080/104000 (14%)] Loss: 0.000010 Train Epoch: 4 [14720/104000 (14%)] Loss: 0.000092 Train Epoch: 4 [15360/104000 (15%)] Loss: 0.000012 Train Epoch: 4 [16000/104000 (15%)] Loss: 0.000026 Train Epoch: 4 [16640/104000 (16%)] Loss: 0.001358 Train Epoch: 4 [17280/104000 (17%)] Loss: 0.000007 Train Epoch: 4 [17920/104000 (17%)] Loss: 0.000006 Train Epoch: 4 [18560/104000 (18%)] Loss: 0.000019 Train Epoch: 4 [19200/104000 (18%)] Loss: 0.000008 Train Epoch: 4 [19840/104000 (19%)] Loss: 0.000014 Train Epoch: 4 [20480/104000 (20%)] Loss: 0.000023 Train Epoch: 4 [21120/104000 (20%)] Loss: 0.000009 Train Epoch: 4 [21760/104000 (21%)] Loss: 0.000013 Train Epoch: 4 [22400/104000 (22%)] Loss: 0.000009 Train Epoch: 4 [23040/104000 (22%)] Loss: 0.000014 Train Epoch: 4 [23680/104000 (23%)] Loss: 0.000013 Train Epoch: 4 [24320/104000 (23%)] Loss: 0.000003 Train Epoch: 4 [24960/104000 (24%)] Loss: 0.000009 Train Epoch: 4 [25600/104000 (25%)] Loss: 0.000012 Train Epoch: 4 [26240/104000 (25%)] Loss: 0.000015 Train Epoch: 4 [26880/104000 (26%)] Loss: 0.000040 Train Epoch: 4 [27520/104000 (26%)] Loss: 0.000005 Train Epoch: 4 [28160/104000 (27%)] Loss: 0.000010 Train Epoch: 4 [28800/104000 (28%)] Loss: 0.000088 Train Epoch: 4 [29440/104000 (28%)] Loss: 0.000006 Train Epoch: 4 [30080/104000 (29%)] Loss: 0.000020 Train Epoch: 4 [30720/104000 (30%)] Loss: 0.000025 Train Epoch: 4 [31360/104000 (30%)] Loss: 0.000025 Train Epoch: 4 [32000/104000 (31%)] Loss: 0.000009 Train Epoch: 4 [32640/104000 (31%)] Loss: 0.000311 Train Epoch: 4 [33280/104000 (32%)] Loss: 0.000019 Train Epoch: 4 [33920/104000 (33%)] Loss: 0.000017 Train Epoch: 4 [34560/104000 (33%)] Loss: 0.000316 Train Epoch: 4 [35200/104000 (34%)] Loss: 0.000131 Train Epoch: 4 [35840/104000 (34%)] Loss: 0.000007 Train Epoch: 4 [36480/104000 (35%)] Loss: 0.000014 Train Epoch: 4 [37120/104000 (36%)] Loss: 0.000032 Train Epoch: 4 [37760/104000 (36%)] Loss: 0.000016 Train Epoch: 4 [38400/104000 (37%)] Loss: 0.000193 Train Epoch: 4 [39040/104000 (38%)] Loss: 0.000029 Train Epoch: 4 [39680/104000 (38%)] Loss: 0.000016 Train Epoch: 4 [40320/104000 (39%)] Loss: 0.000020 Train Epoch: 4 [40960/104000 (39%)] Loss: 0.000008 Train Epoch: 4 [41600/104000 (40%)] Loss: 0.000033 Train Epoch: 4 [42240/104000 (41%)] Loss: 0.001253 Train Epoch: 4 [42880/104000 (41%)] Loss: 0.000018 Train Epoch: 4 [43520/104000 (42%)] Loss: 0.000051 Train Epoch: 4 [44160/104000 (42%)] Loss: 0.000005 Train Epoch: 4 [44800/104000 (43%)] Loss: 0.000011 Train Epoch: 4 [45440/104000 (44%)] Loss: 0.000654 Train Epoch: 4 [46080/104000 (44%)] Loss: 0.000014 Train Epoch: 4 [46720/104000 (45%)] Loss: 0.000011 Train Epoch: 4 [47360/104000 (46%)] Loss: 0.000005 Train Epoch: 4 [48000/104000 (46%)] Loss: 0.000018 Train Epoch: 4 [48640/104000 (47%)] Loss: 0.000007 Train Epoch: 4 [49280/104000 (47%)] Loss: 0.000016 Train Epoch: 4 [49920/104000 (48%)] Loss: 0.000006 Train Epoch: 4 [50560/104000 (49%)] Loss: 0.000016 Train Epoch: 4 [51200/104000 (49%)] Loss: 0.000064 Train Epoch: 4 [51840/104000 (50%)] Loss: 0.000010 Train Epoch: 4 [52480/104000 (50%)] Loss: 0.000099 Train Epoch: 4 [53120/104000 (51%)] Loss: 0.000012 Train Epoch: 4 [53760/104000 (52%)] Loss: 0.000008 Train Epoch: 4 [54400/104000 (52%)] Loss: 0.000005 Train Epoch: 4 [55040/104000 (53%)] Loss: 0.000014 Train Epoch: 4 [55680/104000 (54%)] Loss: 0.000014 Train Epoch: 4 [56320/104000 (54%)] Loss: 0.000025 Train Epoch: 4 [56960/104000 (55%)] Loss: 0.000009 Train Epoch: 4 [57600/104000 (55%)] Loss: 0.000051 Train Epoch: 4 [58240/104000 (56%)] Loss: 0.000021 Train Epoch: 4 [58880/104000 (57%)] Loss: 0.000023 Train Epoch: 4 [59520/104000 (57%)] Loss: 0.000148 Train Epoch: 4 [60160/104000 (58%)] Loss: 0.000017 Train Epoch: 4 [60800/104000 (58%)] Loss: 0.000008 Train Epoch: 4 [61440/104000 (59%)] Loss: 0.000024 Train Epoch: 4 [62080/104000 (60%)] Loss: 0.000053 Train Epoch: 4 [62720/104000 (60%)] Loss: 0.000018 Train Epoch: 4 [63360/104000 (61%)] Loss: 0.000008 Train Epoch: 4 [64000/104000 (62%)] Loss: 0.000004 Train Epoch: 4 [64640/104000 (62%)] Loss: 0.000016 Train Epoch: 4 [65280/104000 (63%)] Loss: 0.000179 Train Epoch: 4 [65920/104000 (63%)] Loss: 0.000004 Train Epoch: 4 [66560/104000 (64%)] Loss: 0.000008 Train Epoch: 4 [67200/104000 (65%)] Loss: 0.000181 Train Epoch: 4 [67840/104000 (65%)] Loss: 0.000023 Train Epoch: 4 [68480/104000 (66%)] Loss: 0.000011 Train Epoch: 4 [69120/104000 (66%)] Loss: 0.000017 Train Epoch: 4 [69760/104000 (67%)] Loss: 0.000010 Train Epoch: 4 [70400/104000 (68%)] Loss: 0.000009 Train Epoch: 4 [71040/104000 (68%)] Loss: 0.000028 Train Epoch: 4 [71680/104000 (69%)] Loss: 0.000007 Train Epoch: 4 [72320/104000 (70%)] Loss: 0.000010 Train Epoch: 4 [72960/104000 (70%)] Loss: 0.000251 Train Epoch: 4 [73600/104000 (71%)] Loss: 0.000011 Train Epoch: 4 [74240/104000 (71%)] Loss: 0.000032 Train Epoch: 4 [74880/104000 (72%)] Loss: 0.000009 Train Epoch: 4 [75520/104000 (73%)] Loss: 0.000013 Train Epoch: 4 [76160/104000 (73%)] Loss: 0.000013 Train Epoch: 4 [76800/104000 (74%)] Loss: 0.000015 Train Epoch: 4 [77440/104000 (74%)] Loss: 0.000022 Train Epoch: 4 [78080/104000 (75%)] Loss: 0.000012 Train Epoch: 4 [78720/104000 (76%)] Loss: 0.000006 Train Epoch: 4 [79360/104000 (76%)] Loss: 0.000086 Train Epoch: 4 [80000/104000 (77%)] Loss: 0.000016 Train Epoch: 4 [80640/104000 (78%)] Loss: 0.000022 Train Epoch: 4 [81280/104000 (78%)] Loss: 0.000047 Train Epoch: 4 [81920/104000 (79%)] Loss: 0.000026 Train Epoch: 4 [82560/104000 (79%)] Loss: 0.000019 Train Epoch: 4 [83200/104000 (80%)] Loss: 0.000036 Train Epoch: 4 [83840/104000 (81%)] Loss: 0.000004 Train Epoch: 4 [84480/104000 (81%)] Loss: 0.000112 Train Epoch: 4 [85120/104000 (82%)] Loss: 0.000016 Train Epoch: 4 [85760/104000 (82%)] Loss: 0.000007 Train Epoch: 4 [86400/104000 (83%)] Loss: 0.000023 Train Epoch: 4 [87040/104000 (84%)] Loss: 0.000045 Train Epoch: 4 [87680/104000 (84%)] Loss: 0.000008 Train Epoch: 4 [88320/104000 (85%)] Loss: 0.000012 Train Epoch: 4 [88960/104000 (86%)] Loss: 0.000013 Train Epoch: 4 [89600/104000 (86%)] Loss: 0.000023 Train Epoch: 4 [90240/104000 (87%)] Loss: 0.000009 Train Epoch: 4 [90880/104000 (87%)] Loss: 0.000028 Train Epoch: 4 [91520/104000 (88%)] Loss: 0.000011 Train Epoch: 4 [92160/104000 (89%)] Loss: 0.000034 Train Epoch: 4 [92800/104000 (89%)] Loss: 0.000008 Train Epoch: 4 [93440/104000 (90%)] Loss: 0.000011 Train Epoch: 4 [94080/104000 (90%)] Loss: 0.000192 Train Epoch: 4 [94720/104000 (91%)] Loss: 0.000012 Train Epoch: 4 [95360/104000 (92%)] Loss: 0.000024 Train Epoch: 4 [96000/104000 (92%)] Loss: 0.000111 Train Epoch: 4 [96640/104000 (93%)] Loss: 0.000016 Train Epoch: 4 [97280/104000 (94%)] Loss: 0.000021 Train Epoch: 4 [97920/104000 (94%)] Loss: 0.000007 Train Epoch: 4 [98560/104000 (95%)] Loss: 0.000025 Train Epoch: 4 [99200/104000 (95%)] Loss: 0.000022 Train Epoch: 4 [99840/104000 (96%)] Loss: 0.000005 Train Epoch: 4 [100480/104000 (97%)] Loss: 0.000006 Train Epoch: 4 [101120/104000 (97%)] Loss: 0.000007 Train Epoch: 4 [101760/104000 (98%)] Loss: 0.000011 Train Epoch: 4 [102400/104000 (98%)] Loss: 0.000016 Train Epoch: 4 [103040/104000 (99%)] Loss: 0.000003 Train Epoch: 4 [103680/104000 (100%)] Loss: 0.000007 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.1927, Accuracy: 27318/30000 (91%) Train Epoch: 5 [0/104000 (0%)] Loss: 0.000032 Train Epoch: 5 [640/104000 (1%)] Loss: 0.000012 Train Epoch: 5 [1280/104000 (1%)] Loss: 0.000008 Train Epoch: 5 [1920/104000 (2%)] Loss: 0.000012 Train Epoch: 5 [2560/104000 (2%)] Loss: 0.000034 Train Epoch: 5 [3200/104000 (3%)] Loss: 0.000011 Train Epoch: 5 [3840/104000 (4%)] Loss: 0.000511 Train Epoch: 5 [4480/104000 (4%)] Loss: 0.000013 Train Epoch: 5 [5120/104000 (5%)] Loss: 0.000025 Train Epoch: 5 [5760/104000 (6%)] Loss: 0.000085 Train Epoch: 5 [6400/104000 (6%)] Loss: 0.000022 Train Epoch: 5 [7040/104000 (7%)] Loss: 0.000012 Train Epoch: 5 [7680/104000 (7%)] Loss: 0.000009 Train Epoch: 5 [8320/104000 (8%)] Loss: 0.000037 Train Epoch: 5 [8960/104000 (9%)] Loss: 0.000015 Train Epoch: 5 [9600/104000 (9%)] Loss: 0.000009 Train Epoch: 5 [10240/104000 (10%)] Loss: 0.000005 Train Epoch: 5 [10880/104000 (10%)] Loss: 0.000008 Train Epoch: 5 [11520/104000 (11%)] Loss: 0.000008 Train Epoch: 5 [12160/104000 (12%)] Loss: 0.000013 Train Epoch: 5 [12800/104000 (12%)] Loss: 0.000008 Train Epoch: 5 [13440/104000 (13%)] Loss: 0.000267 Train Epoch: 5 [14080/104000 (14%)] Loss: 0.000008 Train Epoch: 5 [14720/104000 (14%)] Loss: 0.000086 Train Epoch: 5 [15360/104000 (15%)] Loss: 0.000017 Train Epoch: 5 [16000/104000 (15%)] Loss: 0.000002 Train Epoch: 5 [16640/104000 (16%)] Loss: 0.000006 Train Epoch: 5 [17280/104000 (17%)] Loss: 0.000010 Train Epoch: 5 [17920/104000 (17%)] Loss: 0.000093 Train Epoch: 5 [18560/104000 (18%)] Loss: 0.000016 Train Epoch: 5 [19200/104000 (18%)] Loss: 0.000021 Train Epoch: 5 [19840/104000 (19%)] Loss: 0.000009 Train Epoch: 5 [20480/104000 (20%)] Loss: 0.000015 Train Epoch: 5 [21120/104000 (20%)] Loss: 0.000011 Train Epoch: 5 [21760/104000 (21%)] Loss: 0.000011 Train Epoch: 5 [22400/104000 (22%)] Loss: 0.000007 Train Epoch: 5 [23040/104000 (22%)] Loss: 0.000007 Train Epoch: 5 [23680/104000 (23%)] Loss: 0.000043 Train Epoch: 5 [24320/104000 (23%)] Loss: 0.000021 Train Epoch: 5 [24960/104000 (24%)] Loss: 0.000037 Train Epoch: 5 [25600/104000 (25%)] Loss: 0.000008 Train Epoch: 5 [26240/104000 (25%)] Loss: 0.000005 Train Epoch: 5 [26880/104000 (26%)] Loss: 0.000007 Train Epoch: 5 [27520/104000 (26%)] Loss: 0.000005 Train Epoch: 5 [28160/104000 (27%)] Loss: 0.000008 Train Epoch: 5 [28800/104000 (28%)] Loss: 0.000009 Train Epoch: 5 [29440/104000 (28%)] Loss: 0.000230 Train Epoch: 5 [30080/104000 (29%)] Loss: 0.000009 Train Epoch: 5 [30720/104000 (30%)] Loss: 0.000009 Train Epoch: 5 [31360/104000 (30%)] Loss: 0.000020 Train Epoch: 5 [32000/104000 (31%)] Loss: 0.000012 Train Epoch: 5 [32640/104000 (31%)] Loss: 0.000021 Train Epoch: 5 [33280/104000 (32%)] Loss: 0.000021 Train Epoch: 5 [33920/104000 (33%)] Loss: 0.000004 Train Epoch: 5 [34560/104000 (33%)] Loss: 0.000009 Train Epoch: 5 [35200/104000 (34%)] Loss: 0.000006 Train Epoch: 5 [35840/104000 (34%)] Loss: 0.000004 Train Epoch: 5 [36480/104000 (35%)] Loss: 0.000016 Train Epoch: 5 [37120/104000 (36%)] Loss: 0.000024 Train Epoch: 5 [37760/104000 (36%)] Loss: 0.000008 Train Epoch: 5 [38400/104000 (37%)] Loss: 0.000023 Train Epoch: 5 [39040/104000 (38%)] Loss: 0.000018 Train Epoch: 5 [39680/104000 (38%)] Loss: 0.000010 Train Epoch: 5 [40320/104000 (39%)] Loss: 0.000038 Train Epoch: 5 [40960/104000 (39%)] Loss: 0.000005 Train Epoch: 5 [41600/104000 (40%)] Loss: 0.000008 Train Epoch: 5 [42240/104000 (41%)] Loss: 0.000023 Train Epoch: 5 [42880/104000 (41%)] Loss: 0.000008 Train Epoch: 5 [43520/104000 (42%)] Loss: 0.000018 Train Epoch: 5 [44160/104000 (42%)] Loss: 0.000009 Train Epoch: 5 [44800/104000 (43%)] Loss: 0.000018 Train Epoch: 5 [45440/104000 (44%)] Loss: 0.000013 Train Epoch: 5 [46080/104000 (44%)] Loss: 0.000010 Train Epoch: 5 [46720/104000 (45%)] Loss: 0.000215 Train Epoch: 5 [47360/104000 (46%)] Loss: 0.000012 Train Epoch: 5 [48000/104000 (46%)] Loss: 0.000015 Train Epoch: 5 [48640/104000 (47%)] Loss: 0.000012 Train Epoch: 5 [49280/104000 (47%)] Loss: 0.000007 Train Epoch: 5 [49920/104000 (48%)] Loss: 0.000006 Train Epoch: 5 [50560/104000 (49%)] Loss: 0.000025 Train Epoch: 5 [51200/104000 (49%)] Loss: 0.000068 Train Epoch: 5 [51840/104000 (50%)] Loss: 0.000014 Train Epoch: 5 [52480/104000 (50%)] Loss: 0.000015 Train Epoch: 5 [53120/104000 (51%)] Loss: 0.000009 Train Epoch: 5 [53760/104000 (52%)] Loss: 0.000012 Train Epoch: 5 [54400/104000 (52%)] Loss: 0.000026 Train Epoch: 5 [55040/104000 (53%)] Loss: 0.000012 Train Epoch: 5 [55680/104000 (54%)] Loss: 0.000070 Train Epoch: 5 [56320/104000 (54%)] Loss: 0.000070 Train Epoch: 5 [56960/104000 (55%)] Loss: 0.000069 Train Epoch: 5 [57600/104000 (55%)] Loss: 0.000009 Train Epoch: 5 [58240/104000 (56%)] Loss: 0.000005 Train Epoch: 5 [58880/104000 (57%)] Loss: 0.000012 Train Epoch: 5 [59520/104000 (57%)] Loss: 0.000022 Train Epoch: 5 [60160/104000 (58%)] Loss: 0.000007 Train Epoch: 5 [60800/104000 (58%)] Loss: 0.000007 Train Epoch: 5 [61440/104000 (59%)] Loss: 0.000076 Train Epoch: 5 [62080/104000 (60%)] Loss: 0.000020 Train Epoch: 5 [62720/104000 (60%)] Loss: 0.000012 Train Epoch: 5 [63360/104000 (61%)] Loss: 0.000086 Train Epoch: 5 [64000/104000 (62%)] Loss: 0.000006 Train Epoch: 5 [64640/104000 (62%)] Loss: 0.000009 Train Epoch: 5 [65280/104000 (63%)] Loss: 0.000005 Train Epoch: 5 [65920/104000 (63%)] Loss: 0.000019 Train Epoch: 5 [66560/104000 (64%)] Loss: 0.000099 Train Epoch: 5 [67200/104000 (65%)] Loss: 0.000127 Train Epoch: 5 [67840/104000 (65%)] Loss: 0.000017 Train Epoch: 5 [68480/104000 (66%)] Loss: 0.000021 Train Epoch: 5 [69120/104000 (66%)] Loss: 0.000006 Train Epoch: 5 [69760/104000 (67%)] Loss: 0.000007 Train Epoch: 5 [70400/104000 (68%)] Loss: 0.000006 Train Epoch: 5 [71040/104000 (68%)] Loss: 0.000022 Train Epoch: 5 [71680/104000 (69%)] Loss: 0.000122 Train Epoch: 5 [72320/104000 (70%)] Loss: 0.000021 Train Epoch: 5 [72960/104000 (70%)] Loss: 0.000012 Train Epoch: 5 [73600/104000 (71%)] Loss: 0.000145 Train Epoch: 5 [74240/104000 (71%)] Loss: 0.000236 Train Epoch: 5 [74880/104000 (72%)] Loss: 0.000004 Train Epoch: 5 [75520/104000 (73%)] Loss: 0.000013 Train Epoch: 5 [76160/104000 (73%)] Loss: 0.000009 Train Epoch: 5 [76800/104000 (74%)] Loss: 0.000038 Train Epoch: 5 [77440/104000 (74%)] Loss: 0.000065 Train Epoch: 5 [78080/104000 (75%)] Loss: 0.000014 Train Epoch: 5 [78720/104000 (76%)] Loss: 0.000017 Train Epoch: 5 [79360/104000 (76%)] Loss: 0.000002 Train Epoch: 5 [80000/104000 (77%)] Loss: 0.000004 Train Epoch: 5 [80640/104000 (78%)] Loss: 0.000010 Train Epoch: 5 [81280/104000 (78%)] Loss: 0.000005 Train Epoch: 5 [81920/104000 (79%)] Loss: 0.000016 Train Epoch: 5 [82560/104000 (79%)] Loss: 0.000016 Train Epoch: 5 [83200/104000 (80%)] Loss: 0.000031 Train Epoch: 5 [83840/104000 (81%)] Loss: 0.000007 Train Epoch: 5 [84480/104000 (81%)] Loss: 0.000036 Train Epoch: 5 [85120/104000 (82%)] Loss: 0.000009 Train Epoch: 5 [85760/104000 (82%)] Loss: 0.000003 Train Epoch: 5 [86400/104000 (83%)] Loss: 0.000023 Train Epoch: 5 [87040/104000 (84%)] Loss: 0.000134 Train Epoch: 5 [87680/104000 (84%)] Loss: 0.000017 Train Epoch: 5 [88320/104000 (85%)] Loss: 0.000004 Train Epoch: 5 [88960/104000 (86%)] Loss: 0.000014 Train Epoch: 5 [89600/104000 (86%)] Loss: 0.000060 Train Epoch: 5 [90240/104000 (87%)] Loss: 0.000010 Train Epoch: 5 [90880/104000 (87%)] Loss: 0.000008 Train Epoch: 5 [91520/104000 (88%)] Loss: 0.000006 Train Epoch: 5 [92160/104000 (89%)] Loss: 0.000013 Train Epoch: 5 [92800/104000 (89%)] Loss: 0.000006 Train Epoch: 5 [93440/104000 (90%)] Loss: 0.000010 Train Epoch: 5 [94080/104000 (90%)] Loss: 0.000060 Train Epoch: 5 [94720/104000 (91%)] Loss: 0.000156 Train Epoch: 5 [95360/104000 (92%)] Loss: 0.000115 Train Epoch: 5 [96000/104000 (92%)] Loss: 0.000048 Train Epoch: 5 [96640/104000 (93%)] Loss: 0.000006 Train Epoch: 5 [97280/104000 (94%)] Loss: 0.000009 Train Epoch: 5 [97920/104000 (94%)] Loss: 0.000012 Train Epoch: 5 [98560/104000 (95%)] Loss: 0.000006 Train Epoch: 5 [99200/104000 (95%)] Loss: 0.000010 Train Epoch: 5 [99840/104000 (96%)] Loss: 0.000007 Train Epoch: 5 [100480/104000 (97%)] Loss: 0.000010 Train Epoch: 5 [101120/104000 (97%)] Loss: 0.000033 Train Epoch: 5 [101760/104000 (98%)] Loss: 0.000037 Train Epoch: 5 [102400/104000 (98%)] Loss: 0.000013 Train Epoch: 5 [103040/104000 (99%)] Loss: 0.000003 Train Epoch: 5 [103680/104000 (100%)] Loss: 0.000123 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2115, Accuracy: 27050/30000 (90%) [100.0, 100.0, 100.0] [88.56, 91.44666666666667, 90.16666666666667] The number of neurons in CNN layer is 20 Train Epoch: 1 [0/104000 (0%)] Loss: 0.838881 Train Epoch: 1 [640/104000 (1%)] Loss: 0.000247 Train Epoch: 1 [1280/104000 (1%)] Loss: 0.000319 Train Epoch: 1 [1920/104000 (2%)] Loss: 0.000412 Train Epoch: 1 [2560/104000 (2%)] Loss: 0.000413 Train Epoch: 1 [3200/104000 (3%)] Loss: 0.000391 Train Epoch: 1 [3840/104000 (4%)] Loss: 0.000342 Train Epoch: 1 [4480/104000 (4%)] Loss: 0.000258 Train Epoch: 1 [5120/104000 (5%)] Loss: 0.000206 Train Epoch: 1 [5760/104000 (6%)] Loss: 0.000161 Train Epoch: 1 [6400/104000 (6%)] Loss: 0.000135 Train Epoch: 1 [7040/104000 (7%)] Loss: 0.000136 Train Epoch: 1 [7680/104000 (7%)] Loss: 0.000528 Train Epoch: 1 [8320/104000 (8%)] Loss: 0.000241 Train Epoch: 1 [8960/104000 (9%)] Loss: 0.000707 Train Epoch: 1 [9600/104000 (9%)] Loss: 0.000771 Train Epoch: 1 [10240/104000 (10%)] Loss: 0.000273 Train Epoch: 1 [10880/104000 (10%)] Loss: 0.000157 Train Epoch: 1 [11520/104000 (11%)] Loss: 0.000274 Train Epoch: 1 [12160/104000 (12%)] Loss: 0.000260 Train Epoch: 1 [12800/104000 (12%)] Loss: 0.000103 Train Epoch: 1 [13440/104000 (13%)] Loss: 0.000208 Train Epoch: 1 [14080/104000 (14%)] Loss: 0.000094 Train Epoch: 1 [14720/104000 (14%)] Loss: 0.000089 Train Epoch: 1 [15360/104000 (15%)] Loss: 0.001090 Train Epoch: 1 [16000/104000 (15%)] Loss: 0.000184 Train Epoch: 1 [16640/104000 (16%)] Loss: 0.000120 Train Epoch: 1 [17280/104000 (17%)] Loss: 0.000158 Train Epoch: 1 [17920/104000 (17%)] Loss: 0.000283 Train Epoch: 1 [18560/104000 (18%)] Loss: 0.001772 Train Epoch: 1 [19200/104000 (18%)] Loss: 0.000266 Train Epoch: 1 [19840/104000 (19%)] Loss: 0.000424 Train Epoch: 1 [20480/104000 (20%)] Loss: 0.000118 Train Epoch: 1 [21120/104000 (20%)] Loss: 0.000107 Train Epoch: 1 [21760/104000 (21%)] Loss: 0.000129 Train Epoch: 1 [22400/104000 (22%)] Loss: 0.000481 Train Epoch: 1 [23040/104000 (22%)] Loss: 0.000154 Train Epoch: 1 [23680/104000 (23%)] Loss: 0.001279 Train Epoch: 1 [24320/104000 (23%)] Loss: 0.001023 Train Epoch: 1 [24960/104000 (24%)] Loss: 0.000088 Train Epoch: 1 [25600/104000 (25%)] Loss: 0.000109 Train Epoch: 1 [26240/104000 (25%)] Loss: 0.000168 Train Epoch: 1 [26880/104000 (26%)] Loss: 0.000079 Train Epoch: 1 [27520/104000 (26%)] Loss: 0.000115 Train Epoch: 1 [28160/104000 (27%)] Loss: 0.000096 Train Epoch: 1 [28800/104000 (28%)] Loss: 0.000067 Train Epoch: 1 [29440/104000 (28%)] Loss: 0.000063 Train Epoch: 1 [30080/104000 (29%)] Loss: 0.000077 Train Epoch: 1 [30720/104000 (30%)] Loss: 0.000138 Train Epoch: 1 [31360/104000 (30%)] Loss: 0.000242 Train Epoch: 1 [32000/104000 (31%)] Loss: 0.000136 Train Epoch: 1 [32640/104000 (31%)] Loss: 0.000498 Train Epoch: 1 [33280/104000 (32%)] Loss: 0.000327 Train Epoch: 1 [33920/104000 (33%)] Loss: 0.000224 Train Epoch: 1 [34560/104000 (33%)] Loss: 0.000677 Train Epoch: 1 [35200/104000 (34%)] Loss: 0.000079 Train Epoch: 1 [35840/104000 (34%)] Loss: 0.000067 Train Epoch: 1 [36480/104000 (35%)] Loss: 0.000086 Train Epoch: 1 [37120/104000 (36%)] Loss: 0.000132 Train Epoch: 1 [37760/104000 (36%)] Loss: 0.000195 Train Epoch: 1 [38400/104000 (37%)] Loss: 0.000085 Train Epoch: 1 [39040/104000 (38%)] Loss: 0.000071 Train Epoch: 1 [39680/104000 (38%)] Loss: 0.000099 Train Epoch: 1 [40320/104000 (39%)] Loss: 0.000136 Train Epoch: 1 [40960/104000 (39%)] Loss: 0.000041 Train Epoch: 1 [41600/104000 (40%)] Loss: 0.000178 Train Epoch: 1 [42240/104000 (41%)] Loss: 0.000059 Train Epoch: 1 [42880/104000 (41%)] Loss: 0.000080 Train Epoch: 1 [43520/104000 (42%)] Loss: 0.000081 Train Epoch: 1 [44160/104000 (42%)] Loss: 0.000036 Train Epoch: 1 [44800/104000 (43%)] Loss: 0.000212 Train Epoch: 1 [45440/104000 (44%)] Loss: 0.000143 Train Epoch: 1 [46080/104000 (44%)] Loss: 0.000042 Train Epoch: 1 [46720/104000 (45%)] Loss: 0.000037 Train Epoch: 1 [47360/104000 (46%)] Loss: 0.000304 Train Epoch: 1 [48000/104000 (46%)] Loss: 0.000308 Train Epoch: 1 [48640/104000 (47%)] Loss: 0.000110 Train Epoch: 1 [49280/104000 (47%)] Loss: 0.000160 Train Epoch: 1 [49920/104000 (48%)] Loss: 0.000273 Train Epoch: 1 [50560/104000 (49%)] Loss: 0.001155 Train Epoch: 1 [51200/104000 (49%)] Loss: 0.000160 Train Epoch: 1 [51840/104000 (50%)] Loss: 0.000068 Train Epoch: 1 [52480/104000 (50%)] Loss: 0.000114 Train Epoch: 1 [53120/104000 (51%)] Loss: 0.000130 Train Epoch: 1 [53760/104000 (52%)] Loss: 0.000039 Train Epoch: 1 [54400/104000 (52%)] Loss: 0.000042 Train Epoch: 1 [55040/104000 (53%)] Loss: 0.000034 Train Epoch: 1 [55680/104000 (54%)] Loss: 0.000041 Train Epoch: 1 [56320/104000 (54%)] Loss: 0.000199 Train Epoch: 1 [56960/104000 (55%)] Loss: 0.000083 Train Epoch: 1 [57600/104000 (55%)] Loss: 0.000072 Train Epoch: 1 [58240/104000 (56%)] Loss: 0.000156 Train Epoch: 1 [58880/104000 (57%)] Loss: 0.000081 Train Epoch: 1 [59520/104000 (57%)] Loss: 0.000045 Train Epoch: 1 [60160/104000 (58%)] Loss: 0.000072 Train Epoch: 1 [60800/104000 (58%)] Loss: 0.000081 Train Epoch: 1 [61440/104000 (59%)] Loss: 0.000658 Train Epoch: 1 [62080/104000 (60%)] Loss: 0.000097 Train Epoch: 1 [62720/104000 (60%)] Loss: 0.000073 Train Epoch: 1 [63360/104000 (61%)] Loss: 0.000036 Train Epoch: 1 [64000/104000 (62%)] Loss: 0.000048 Train Epoch: 1 [64640/104000 (62%)] Loss: 0.000230 Train Epoch: 1 [65280/104000 (63%)] Loss: 0.000046 Train Epoch: 1 [65920/104000 (63%)] Loss: 0.000088 Train Epoch: 1 [66560/104000 (64%)] Loss: 0.000138 Train Epoch: 1 [67200/104000 (65%)] Loss: 0.000102 Train Epoch: 1 [67840/104000 (65%)] Loss: 0.000067 Train Epoch: 1 [68480/104000 (66%)] Loss: 0.000061 Train Epoch: 1 [69120/104000 (66%)] Loss: 0.000054 Train Epoch: 1 [69760/104000 (67%)] Loss: 0.000094 Train Epoch: 1 [70400/104000 (68%)] Loss: 0.000031 Train Epoch: 1 [71040/104000 (68%)] Loss: 0.000020 Train Epoch: 1 [71680/104000 (69%)] Loss: 0.000097 Train Epoch: 1 [72320/104000 (70%)] Loss: 0.000029 Train Epoch: 1 [72960/104000 (70%)] Loss: 0.000068 Train Epoch: 1 [73600/104000 (71%)] Loss: 0.000026 Train Epoch: 1 [74240/104000 (71%)] Loss: 0.000280 Train Epoch: 1 [74880/104000 (72%)] Loss: 0.000192 Train Epoch: 1 [75520/104000 (73%)] Loss: 0.000072 Train Epoch: 1 [76160/104000 (73%)] Loss: 0.000275 Train Epoch: 1 [76800/104000 (74%)] Loss: 0.000050 Train Epoch: 1 [77440/104000 (74%)] Loss: 0.000067 Train Epoch: 1 [78080/104000 (75%)] Loss: 0.000064 Train Epoch: 1 [78720/104000 (76%)] Loss: 0.000075 Train Epoch: 1 [79360/104000 (76%)] Loss: 0.000049 Train Epoch: 1 [80000/104000 (77%)] Loss: 0.000106 Train Epoch: 1 [80640/104000 (78%)] Loss: 0.000031 Train Epoch: 1 [81280/104000 (78%)] Loss: 0.000327 Train Epoch: 1 [81920/104000 (79%)] Loss: 0.000087 Train Epoch: 1 [82560/104000 (79%)] Loss: 0.000025 Train Epoch: 1 [83200/104000 (80%)] Loss: 0.000041 Train Epoch: 1 [83840/104000 (81%)] Loss: 0.000026 Train Epoch: 1 [84480/104000 (81%)] Loss: 0.000110 Train Epoch: 1 [85120/104000 (82%)] Loss: 0.000016 Train Epoch: 1 [85760/104000 (82%)] Loss: 0.000031 Train Epoch: 1 [86400/104000 (83%)] Loss: 0.000027 Train Epoch: 1 [87040/104000 (84%)] Loss: 0.000044 Train Epoch: 1 [87680/104000 (84%)] Loss: 0.000048 Train Epoch: 1 [88320/104000 (85%)] Loss: 0.000054 Train Epoch: 1 [88960/104000 (86%)] Loss: 0.000034 Train Epoch: 1 [89600/104000 (86%)] Loss: 0.000043 Train Epoch: 1 [90240/104000 (87%)] Loss: 0.000054 Train Epoch: 1 [90880/104000 (87%)] Loss: 0.000130 Train Epoch: 1 [91520/104000 (88%)] Loss: 0.000110 Train Epoch: 1 [92160/104000 (89%)] Loss: 0.000069 Train Epoch: 1 [92800/104000 (89%)] Loss: 0.000036 Train Epoch: 1 [93440/104000 (90%)] Loss: 0.000028 Train Epoch: 1 [94080/104000 (90%)] Loss: 0.000022 Train Epoch: 1 [94720/104000 (91%)] Loss: 0.000058 Train Epoch: 1 [95360/104000 (92%)] Loss: 0.000016 Train Epoch: 1 [96000/104000 (92%)] Loss: 0.000045 Train Epoch: 1 [96640/104000 (93%)] Loss: 0.000174 Train Epoch: 1 [97280/104000 (94%)] Loss: 0.000091 Train Epoch: 1 [97920/104000 (94%)] Loss: 0.000419 Train Epoch: 1 [98560/104000 (95%)] Loss: 0.000254 Train Epoch: 1 [99200/104000 (95%)] Loss: 0.000016 Train Epoch: 1 [99840/104000 (96%)] Loss: 0.000049 Train Epoch: 1 [100480/104000 (97%)] Loss: 0.000031 Train Epoch: 1 [101120/104000 (97%)] Loss: 0.006004 Train Epoch: 1 [101760/104000 (98%)] Loss: 0.000035 Train Epoch: 1 [102400/104000 (98%)] Loss: 0.000055 Train Epoch: 1 [103040/104000 (99%)] Loss: 0.000103 Train Epoch: 1 [103680/104000 (100%)] Loss: 0.000018 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2020, Accuracy: 27074/30000 (90%) Train Epoch: 2 [0/104000 (0%)] Loss: 0.000029 Train Epoch: 2 [640/104000 (1%)] Loss: 0.000105 Train Epoch: 2 [1280/104000 (1%)] Loss: 0.000028 Train Epoch: 2 [1920/104000 (2%)] Loss: 0.000515 Train Epoch: 2 [2560/104000 (2%)] Loss: 0.000029 Train Epoch: 2 [3200/104000 (3%)] Loss: 0.000050 Train Epoch: 2 [3840/104000 (4%)] Loss: 0.000240 Train Epoch: 2 [4480/104000 (4%)] Loss: 0.000127 Train Epoch: 2 [5120/104000 (5%)] Loss: 0.000032 Train Epoch: 2 [5760/104000 (6%)] Loss: 0.000113 Train Epoch: 2 [6400/104000 (6%)] Loss: 0.000029 Train Epoch: 2 [7040/104000 (7%)] Loss: 0.000015 Train Epoch: 2 [7680/104000 (7%)] Loss: 0.000020 Train Epoch: 2 [8320/104000 (8%)] Loss: 0.000091 Train Epoch: 2 [8960/104000 (9%)] Loss: 0.000032 Train Epoch: 2 [9600/104000 (9%)] Loss: 0.000016 Train Epoch: 2 [10240/104000 (10%)] Loss: 0.000034 Train Epoch: 2 [10880/104000 (10%)] Loss: 0.000012 Train Epoch: 2 [11520/104000 (11%)] Loss: 0.000044 Train Epoch: 2 [12160/104000 (12%)] Loss: 0.000037 Train Epoch: 2 [12800/104000 (12%)] Loss: 0.000054 Train Epoch: 2 [13440/104000 (13%)] Loss: 0.000017 Train Epoch: 2 [14080/104000 (14%)] Loss: 0.000048 Train Epoch: 2 [14720/104000 (14%)] Loss: 0.000028 Train Epoch: 2 [15360/104000 (15%)] Loss: 0.000021 Train Epoch: 2 [16000/104000 (15%)] Loss: 0.000016 Train Epoch: 2 [16640/104000 (16%)] Loss: 0.000027 Train Epoch: 2 [17280/104000 (17%)] Loss: 0.000095 Train Epoch: 2 [17920/104000 (17%)] Loss: 0.000018 Train Epoch: 2 [18560/104000 (18%)] Loss: 0.000121 Train Epoch: 2 [19200/104000 (18%)] Loss: 0.000019 Train Epoch: 2 [19840/104000 (19%)] Loss: 0.000008 Train Epoch: 2 [20480/104000 (20%)] Loss: 0.000017 Train Epoch: 2 [21120/104000 (20%)] Loss: 0.000089 Train Epoch: 2 [21760/104000 (21%)] Loss: 0.000056 Train Epoch: 2 [22400/104000 (22%)] Loss: 0.000010 Train Epoch: 2 [23040/104000 (22%)] Loss: 0.000024 Train Epoch: 2 [23680/104000 (23%)] Loss: 0.000093 Train Epoch: 2 [24320/104000 (23%)] Loss: 0.000022 Train Epoch: 2 [24960/104000 (24%)] Loss: 0.000027 Train Epoch: 2 [25600/104000 (25%)] Loss: 0.000089 Train Epoch: 2 [26240/104000 (25%)] Loss: 0.000019 Train Epoch: 2 [26880/104000 (26%)] Loss: 0.000258 Train Epoch: 2 [27520/104000 (26%)] Loss: 0.000018 Train Epoch: 2 [28160/104000 (27%)] Loss: 0.000007 Train Epoch: 2 [28800/104000 (28%)] Loss: 0.000018 Train Epoch: 2 [29440/104000 (28%)] Loss: 0.000057 Train Epoch: 2 [30080/104000 (29%)] Loss: 0.000011 Train Epoch: 2 [30720/104000 (30%)] Loss: 0.000012 Train Epoch: 2 [31360/104000 (30%)] Loss: 0.000017 Train Epoch: 2 [32000/104000 (31%)] Loss: 0.000051 Train Epoch: 2 [32640/104000 (31%)] Loss: 0.000011 Train Epoch: 2 [33280/104000 (32%)] Loss: 0.000019 Train Epoch: 2 [33920/104000 (33%)] Loss: 0.000753 Train Epoch: 2 [34560/104000 (33%)] Loss: 0.000037 Train Epoch: 2 [35200/104000 (34%)] Loss: 0.000026 Train Epoch: 2 [35840/104000 (34%)] Loss: 0.000099 Train Epoch: 2 [36480/104000 (35%)] Loss: 0.000037 Train Epoch: 2 [37120/104000 (36%)] Loss: 0.000019 Train Epoch: 2 [37760/104000 (36%)] Loss: 0.000008 Train Epoch: 2 [38400/104000 (37%)] Loss: 0.000019 Train Epoch: 2 [39040/104000 (38%)] Loss: 0.000011 Train Epoch: 2 [39680/104000 (38%)] Loss: 0.000110 Train Epoch: 2 [40320/104000 (39%)] Loss: 0.000026 Train Epoch: 2 [40960/104000 (39%)] Loss: 0.000008 Train Epoch: 2 [41600/104000 (40%)] Loss: 0.000137 Train Epoch: 2 [42240/104000 (41%)] Loss: 0.000023 Train Epoch: 2 [42880/104000 (41%)] Loss: 0.000011 Train Epoch: 2 [43520/104000 (42%)] Loss: 0.000045 Train Epoch: 2 [44160/104000 (42%)] Loss: 0.000020 Train Epoch: 2 [44800/104000 (43%)] Loss: 0.000013 Train Epoch: 2 [45440/104000 (44%)] Loss: 0.000048 Train Epoch: 2 [46080/104000 (44%)] Loss: 0.000009 Train Epoch: 2 [46720/104000 (45%)] Loss: 0.000013 Train Epoch: 2 [47360/104000 (46%)] Loss: 0.000101 Train Epoch: 2 [48000/104000 (46%)] Loss: 0.000010 Train Epoch: 2 [48640/104000 (47%)] Loss: 0.000063 Train Epoch: 2 [49280/104000 (47%)] Loss: 0.000170 Train Epoch: 2 [49920/104000 (48%)] Loss: 0.000005 Train Epoch: 2 [50560/104000 (49%)] Loss: 0.000034 Train Epoch: 2 [51200/104000 (49%)] Loss: 0.000011 Train Epoch: 2 [51840/104000 (50%)] Loss: 0.000045 Train Epoch: 2 [52480/104000 (50%)] Loss: 0.000013 Train Epoch: 2 [53120/104000 (51%)] Loss: 0.000018 Train Epoch: 2 [53760/104000 (52%)] Loss: 0.000685 Train Epoch: 2 [54400/104000 (52%)] Loss: 0.000035 Train Epoch: 2 [55040/104000 (53%)] Loss: 0.000171 Train Epoch: 2 [55680/104000 (54%)] Loss: 0.000015 Train Epoch: 2 [56320/104000 (54%)] Loss: 0.000009 Train Epoch: 2 [56960/104000 (55%)] Loss: 0.000020 Train Epoch: 2 [57600/104000 (55%)] Loss: 0.000031 Train Epoch: 2 [58240/104000 (56%)] Loss: 0.000010 Train Epoch: 2 [58880/104000 (57%)] Loss: 0.000042 Train Epoch: 2 [59520/104000 (57%)] Loss: 0.000016 Train Epoch: 2 [60160/104000 (58%)] Loss: 0.000031 Train Epoch: 2 [60800/104000 (58%)] Loss: 0.000074 Train Epoch: 2 [61440/104000 (59%)] Loss: 0.000042 Train Epoch: 2 [62080/104000 (60%)] Loss: 0.000018 Train Epoch: 2 [62720/104000 (60%)] Loss: 0.000044 Train Epoch: 2 [63360/104000 (61%)] Loss: 0.000040 Train Epoch: 2 [64000/104000 (62%)] Loss: 0.000017 Train Epoch: 2 [64640/104000 (62%)] Loss: 0.000026 Train Epoch: 2 [65280/104000 (63%)] Loss: 0.000050 Train Epoch: 2 [65920/104000 (63%)] Loss: 0.000022 Train Epoch: 2 [66560/104000 (64%)] Loss: 0.000013 Train Epoch: 2 [67200/104000 (65%)] Loss: 0.000011 Train Epoch: 2 [67840/104000 (65%)] Loss: 0.000015 Train Epoch: 2 [68480/104000 (66%)] Loss: 0.000009 Train Epoch: 2 [69120/104000 (66%)] Loss: 0.000185 Train Epoch: 2 [69760/104000 (67%)] Loss: 0.000013 Train Epoch: 2 [70400/104000 (68%)] Loss: 0.000013 Train Epoch: 2 [71040/104000 (68%)] Loss: 0.000051 Train Epoch: 2 [71680/104000 (69%)] Loss: 0.000679 Train Epoch: 2 [72320/104000 (70%)] Loss: 0.000024 Train Epoch: 2 [72960/104000 (70%)] Loss: 0.000019 Train Epoch: 2 [73600/104000 (71%)] Loss: 0.000051 Train Epoch: 2 [74240/104000 (71%)] Loss: 0.000061 Train Epoch: 2 [74880/104000 (72%)] Loss: 0.000019 Train Epoch: 2 [75520/104000 (73%)] Loss: 0.000027 Train Epoch: 2 [76160/104000 (73%)] Loss: 0.000077 Train Epoch: 2 [76800/104000 (74%)] Loss: 0.000009 Train Epoch: 2 [77440/104000 (74%)] Loss: 0.000010 Train Epoch: 2 [78080/104000 (75%)] Loss: 0.000012 Train Epoch: 2 [78720/104000 (76%)] Loss: 0.000013 Train Epoch: 2 [79360/104000 (76%)] Loss: 0.000029 Train Epoch: 2 [80000/104000 (77%)] Loss: 0.000015 Train Epoch: 2 [80640/104000 (78%)] Loss: 0.000034 Train Epoch: 2 [81280/104000 (78%)] Loss: 0.000028 Train Epoch: 2 [81920/104000 (79%)] Loss: 0.000027 Train Epoch: 2 [82560/104000 (79%)] Loss: 0.000008 Train Epoch: 2 [83200/104000 (80%)] Loss: 0.000008 Train Epoch: 2 [83840/104000 (81%)] Loss: 0.000009 Train Epoch: 2 [84480/104000 (81%)] Loss: 0.000221 Train Epoch: 2 [85120/104000 (82%)] Loss: 0.000295 Train Epoch: 2 [85760/104000 (82%)] Loss: 0.000021 Train Epoch: 2 [86400/104000 (83%)] Loss: 0.000078 Train Epoch: 2 [87040/104000 (84%)] Loss: 0.000030 Train Epoch: 2 [87680/104000 (84%)] Loss: 0.000041 Train Epoch: 2 [88320/104000 (85%)] Loss: 0.000014 Train Epoch: 2 [88960/104000 (86%)] Loss: 0.000192 Train Epoch: 2 [89600/104000 (86%)] Loss: 0.000028 Train Epoch: 2 [90240/104000 (87%)] Loss: 0.000109 Train Epoch: 2 [90880/104000 (87%)] Loss: 0.000029 Train Epoch: 2 [91520/104000 (88%)] Loss: 0.000008 Train Epoch: 2 [92160/104000 (89%)] Loss: 0.000020 Train Epoch: 2 [92800/104000 (89%)] Loss: 0.000016 Train Epoch: 2 [93440/104000 (90%)] Loss: 0.000058 Train Epoch: 2 [94080/104000 (90%)] Loss: 0.000028 Train Epoch: 2 [94720/104000 (91%)] Loss: 0.000011 Train Epoch: 2 [95360/104000 (92%)] Loss: 0.000010 Train Epoch: 2 [96000/104000 (92%)] Loss: 0.000033 Train Epoch: 2 [96640/104000 (93%)] Loss: 0.000015 Train Epoch: 2 [97280/104000 (94%)] Loss: 0.000029 Train Epoch: 2 [97920/104000 (94%)] Loss: 0.000041 Train Epoch: 2 [98560/104000 (95%)] Loss: 0.000035 Train Epoch: 2 [99200/104000 (95%)] Loss: 0.000008 Train Epoch: 2 [99840/104000 (96%)] Loss: 0.000015 Train Epoch: 2 [100480/104000 (97%)] Loss: 0.000085 Train Epoch: 2 [101120/104000 (97%)] Loss: 0.000006 Train Epoch: 2 [101760/104000 (98%)] Loss: 0.000015 Train Epoch: 2 [102400/104000 (98%)] Loss: 0.000065 Train Epoch: 2 [103040/104000 (99%)] Loss: 0.000008 Train Epoch: 2 [103680/104000 (100%)] Loss: 0.000034 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2114, Accuracy: 26974/30000 (90%) Train Epoch: 3 [0/104000 (0%)] Loss: 0.000035 Train Epoch: 3 [640/104000 (1%)] Loss: 0.000016 Train Epoch: 3 [1280/104000 (1%)] Loss: 0.000010 Train Epoch: 3 [1920/104000 (2%)] Loss: 0.000018 Train Epoch: 3 [2560/104000 (2%)] Loss: 0.000031 Train Epoch: 3 [3200/104000 (3%)] Loss: 0.000017 Train Epoch: 3 [3840/104000 (4%)] Loss: 0.000069 Train Epoch: 3 [4480/104000 (4%)] Loss: 0.000030 Train Epoch: 3 [5120/104000 (5%)] Loss: 0.000021 Train Epoch: 3 [5760/104000 (6%)] Loss: 0.000013 Train Epoch: 3 [6400/104000 (6%)] Loss: 0.000048 Train Epoch: 3 [7040/104000 (7%)] Loss: 0.000009 Train Epoch: 3 [7680/104000 (7%)] Loss: 0.000009 Train Epoch: 3 [8320/104000 (8%)] Loss: 0.000017 Train Epoch: 3 [8960/104000 (9%)] Loss: 0.000008 Train Epoch: 3 [9600/104000 (9%)] Loss: 0.000014 Train Epoch: 3 [10240/104000 (10%)] Loss: 0.000019 Train Epoch: 3 [10880/104000 (10%)] Loss: 0.000035 Train Epoch: 3 [11520/104000 (11%)] Loss: 0.000179 Train Epoch: 3 [12160/104000 (12%)] Loss: 0.000009 Train Epoch: 3 [12800/104000 (12%)] Loss: 0.000013 Train Epoch: 3 [13440/104000 (13%)] Loss: 0.000016 Train Epoch: 3 [14080/104000 (14%)] Loss: 0.000052 Train Epoch: 3 [14720/104000 (14%)] Loss: 0.000011 Train Epoch: 3 [15360/104000 (15%)] Loss: 0.000129 Train Epoch: 3 [16000/104000 (15%)] Loss: 0.000133 Train Epoch: 3 [16640/104000 (16%)] Loss: 0.000011 Train Epoch: 3 [17280/104000 (17%)] Loss: 0.000008 Train Epoch: 3 [17920/104000 (17%)] Loss: 0.000282 Train Epoch: 3 [18560/104000 (18%)] Loss: 0.000007 Train Epoch: 3 [19200/104000 (18%)] Loss: 0.000007 Train Epoch: 3 [19840/104000 (19%)] Loss: 0.000020 Train Epoch: 3 [20480/104000 (20%)] Loss: 0.000005 Train Epoch: 3 [21120/104000 (20%)] Loss: 0.000010 Train Epoch: 3 [21760/104000 (21%)] Loss: 0.000008 Train Epoch: 3 [22400/104000 (22%)] Loss: 0.000017 Train Epoch: 3 [23040/104000 (22%)] Loss: 0.000009 Train Epoch: 3 [23680/104000 (23%)] Loss: 0.000065 Train Epoch: 3 [24320/104000 (23%)] Loss: 0.000013 Train Epoch: 3 [24960/104000 (24%)] Loss: 0.000007 Train Epoch: 3 [25600/104000 (25%)] Loss: 0.000021 Train Epoch: 3 [26240/104000 (25%)] Loss: 0.000007 Train Epoch: 3 [26880/104000 (26%)] Loss: 0.000016 Train Epoch: 3 [27520/104000 (26%)] Loss: 0.000013 Train Epoch: 3 [28160/104000 (27%)] Loss: 0.000008 Train Epoch: 3 [28800/104000 (28%)] Loss: 0.000011 Train Epoch: 3 [29440/104000 (28%)] Loss: 0.000034 Train Epoch: 3 [30080/104000 (29%)] Loss: 0.000070 Train Epoch: 3 [30720/104000 (30%)] Loss: 0.000006 Train Epoch: 3 [31360/104000 (30%)] Loss: 0.000009 Train Epoch: 3 [32000/104000 (31%)] Loss: 0.000007 Train Epoch: 3 [32640/104000 (31%)] Loss: 0.000016 Train Epoch: 3 [33280/104000 (32%)] Loss: 0.000020 Train Epoch: 3 [33920/104000 (33%)] Loss: 0.000021 Train Epoch: 3 [34560/104000 (33%)] Loss: 0.000019 Train Epoch: 3 [35200/104000 (34%)] Loss: 0.000008 Train Epoch: 3 [35840/104000 (34%)] Loss: 0.000005 Train Epoch: 3 [36480/104000 (35%)] Loss: 0.000010 Train Epoch: 3 [37120/104000 (36%)] Loss: 0.000046 Train Epoch: 3 [37760/104000 (36%)] Loss: 0.000043 Train Epoch: 3 [38400/104000 (37%)] Loss: 0.000015 Train Epoch: 3 [39040/104000 (38%)] Loss: 0.000008 Train Epoch: 3 [39680/104000 (38%)] Loss: 0.000008 Train Epoch: 3 [40320/104000 (39%)] Loss: 0.000031 Train Epoch: 3 [40960/104000 (39%)] Loss: 0.000158 Train Epoch: 3 [41600/104000 (40%)] Loss: 0.000085 Train Epoch: 3 [42240/104000 (41%)] Loss: 0.000009 Train Epoch: 3 [42880/104000 (41%)] Loss: 0.000013 Train Epoch: 3 [43520/104000 (42%)] Loss: 0.000014 Train Epoch: 3 [44160/104000 (42%)] Loss: 0.000016 Train Epoch: 3 [44800/104000 (43%)] Loss: 0.000107 Train Epoch: 3 [45440/104000 (44%)] Loss: 0.000022 Train Epoch: 3 [46080/104000 (44%)] Loss: 0.000008 Train Epoch: 3 [46720/104000 (45%)] Loss: 0.000007 Train Epoch: 3 [47360/104000 (46%)] Loss: 0.000066 Train Epoch: 3 [48000/104000 (46%)] Loss: 0.000015 Train Epoch: 3 [48640/104000 (47%)] Loss: 0.000012 Train Epoch: 3 [49280/104000 (47%)] Loss: 0.000007 Train Epoch: 3 [49920/104000 (48%)] Loss: 0.000021 Train Epoch: 3 [50560/104000 (49%)] Loss: 0.000006 Train Epoch: 3 [51200/104000 (49%)] Loss: 0.000009 Train Epoch: 3 [51840/104000 (50%)] Loss: 0.000077 Train Epoch: 3 [52480/104000 (50%)] Loss: 0.000013 Train Epoch: 3 [53120/104000 (51%)] Loss: 0.000008 Train Epoch: 3 [53760/104000 (52%)] Loss: 0.000008 Train Epoch: 3 [54400/104000 (52%)] Loss: 0.000018 Train Epoch: 3 [55040/104000 (53%)] Loss: 0.000012 Train Epoch: 3 [55680/104000 (54%)] Loss: 0.000075 Train Epoch: 3 [56320/104000 (54%)] Loss: 0.000010 Train Epoch: 3 [56960/104000 (55%)] Loss: 0.000018 Train Epoch: 3 [57600/104000 (55%)] Loss: 0.000010 Train Epoch: 3 [58240/104000 (56%)] Loss: 0.000127 Train Epoch: 3 [58880/104000 (57%)] Loss: 0.000006 Train Epoch: 3 [59520/104000 (57%)] Loss: 0.000004 Train Epoch: 3 [60160/104000 (58%)] Loss: 0.000019 Train Epoch: 3 [60800/104000 (58%)] Loss: 0.000042 Train Epoch: 3 [61440/104000 (59%)] Loss: 0.000031 Train Epoch: 3 [62080/104000 (60%)] Loss: 0.000004 Train Epoch: 3 [62720/104000 (60%)] Loss: 0.000119 Train Epoch: 3 [63360/104000 (61%)] Loss: 0.000048 Train Epoch: 3 [64000/104000 (62%)] Loss: 0.000023 Train Epoch: 3 [64640/104000 (62%)] Loss: 0.000011 Train Epoch: 3 [65280/104000 (63%)] Loss: 0.000023 Train Epoch: 3 [65920/104000 (63%)] Loss: 0.000021 Train Epoch: 3 [66560/104000 (64%)] Loss: 0.000063 Train Epoch: 3 [67200/104000 (65%)] Loss: 0.000072 Train Epoch: 3 [67840/104000 (65%)] Loss: 0.000012 Train Epoch: 3 [68480/104000 (66%)] Loss: 0.000014 Train Epoch: 3 [69120/104000 (66%)] Loss: 0.000016 Train Epoch: 3 [69760/104000 (67%)] Loss: 0.000072 Train Epoch: 3 [70400/104000 (68%)] Loss: 0.000012 Train Epoch: 3 [71040/104000 (68%)] Loss: 0.000011 Train Epoch: 3 [71680/104000 (69%)] Loss: 0.000021 Train Epoch: 3 [72320/104000 (70%)] Loss: 0.000011 Train Epoch: 3 [72960/104000 (70%)] Loss: 0.000006 Train Epoch: 3 [73600/104000 (71%)] Loss: 0.000006 Train Epoch: 3 [74240/104000 (71%)] Loss: 0.000035 Train Epoch: 3 [74880/104000 (72%)] Loss: 0.000010 Train Epoch: 3 [75520/104000 (73%)] Loss: 0.000009 Train Epoch: 3 [76160/104000 (73%)] Loss: 0.000009 Train Epoch: 3 [76800/104000 (74%)] Loss: 0.000086 Train Epoch: 3 [77440/104000 (74%)] Loss: 0.000047 Train Epoch: 3 [78080/104000 (75%)] Loss: 0.000012 Train Epoch: 3 [78720/104000 (76%)] Loss: 0.000030 Train Epoch: 3 [79360/104000 (76%)] Loss: 0.000054 Train Epoch: 3 [80000/104000 (77%)] Loss: 0.000047 Train Epoch: 3 [80640/104000 (78%)] Loss: 0.000015 Train Epoch: 3 [81280/104000 (78%)] Loss: 0.000043 Train Epoch: 3 [81920/104000 (79%)] Loss: 0.000006 Train Epoch: 3 [82560/104000 (79%)] Loss: 0.000030 Train Epoch: 3 [83200/104000 (80%)] Loss: 0.000012 Train Epoch: 3 [83840/104000 (81%)] Loss: 0.000007 Train Epoch: 3 [84480/104000 (81%)] Loss: 0.000006 Train Epoch: 3 [85120/104000 (82%)] Loss: 0.000010 Train Epoch: 3 [85760/104000 (82%)] Loss: 0.000011 Train Epoch: 3 [86400/104000 (83%)] Loss: 0.000014 Train Epoch: 3 [87040/104000 (84%)] Loss: 0.000008 Train Epoch: 3 [87680/104000 (84%)] Loss: 0.000024 Train Epoch: 3 [88320/104000 (85%)] Loss: 0.000018 Train Epoch: 3 [88960/104000 (86%)] Loss: 0.000089 Train Epoch: 3 [89600/104000 (86%)] Loss: 0.000048 Train Epoch: 3 [90240/104000 (87%)] Loss: 0.000016 Train Epoch: 3 [90880/104000 (87%)] Loss: 0.000011 Train Epoch: 3 [91520/104000 (88%)] Loss: 0.000006 Train Epoch: 3 [92160/104000 (89%)] Loss: 0.000028 Train Epoch: 3 [92800/104000 (89%)] Loss: 0.000079 Train Epoch: 3 [93440/104000 (90%)] Loss: 0.000007 Train Epoch: 3 [94080/104000 (90%)] Loss: 0.000007 Train Epoch: 3 [94720/104000 (91%)] Loss: 0.000003 Train Epoch: 3 [95360/104000 (92%)] Loss: 0.000012 Train Epoch: 3 [96000/104000 (92%)] Loss: 0.000007 Train Epoch: 3 [96640/104000 (93%)] Loss: 0.000005 Train Epoch: 3 [97280/104000 (94%)] Loss: 0.000006 Train Epoch: 3 [97920/104000 (94%)] Loss: 0.000008 Train Epoch: 3 [98560/104000 (95%)] Loss: 0.000005 Train Epoch: 3 [99200/104000 (95%)] Loss: 0.000012 Train Epoch: 3 [99840/104000 (96%)] Loss: 0.000010 Train Epoch: 3 [100480/104000 (97%)] Loss: 0.000010 Train Epoch: 3 [101120/104000 (97%)] Loss: 0.000023 Train Epoch: 3 [101760/104000 (98%)] Loss: 0.000023 Train Epoch: 3 [102400/104000 (98%)] Loss: 0.000019 Train Epoch: 3 [103040/104000 (99%)] Loss: 0.000008 Train Epoch: 3 [103680/104000 (100%)] Loss: 0.000052 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.2591, Accuracy: 26312/30000 (88%) Train Epoch: 4 [0/104000 (0%)] Loss: 0.000011 Train Epoch: 4 [640/104000 (1%)] Loss: 0.000014 Train Epoch: 4 [1280/104000 (1%)] Loss: 0.000014 Train Epoch: 4 [1920/104000 (2%)] Loss: 0.000007 Train Epoch: 4 [2560/104000 (2%)] Loss: 0.000013 Train Epoch: 4 [3200/104000 (3%)] Loss: 0.000009 Train Epoch: 4 [3840/104000 (4%)] Loss: 0.000006 Train Epoch: 4 [4480/104000 (4%)] Loss: 0.000232 Train Epoch: 4 [5120/104000 (5%)] Loss: 0.000019 Train Epoch: 4 [5760/104000 (6%)] Loss: 0.000023 Train Epoch: 4 [6400/104000 (6%)] Loss: 0.000005 Train Epoch: 4 [7040/104000 (7%)] Loss: 0.000014 Train Epoch: 4 [7680/104000 (7%)] Loss: 0.000010 Train Epoch: 4 [8320/104000 (8%)] Loss: 0.000088 Train Epoch: 4 [8960/104000 (9%)] Loss: 0.000037 Train Epoch: 4 [9600/104000 (9%)] Loss: 0.000008 Train Epoch: 4 [10240/104000 (10%)] Loss: 0.000023 Train Epoch: 4 [10880/104000 (10%)] Loss: 0.000021 Train Epoch: 4 [11520/104000 (11%)] Loss: 0.000008 Train Epoch: 4 [12160/104000 (12%)] Loss: 0.000062 Train Epoch: 4 [12800/104000 (12%)] Loss: 0.000174 Train Epoch: 4 [13440/104000 (13%)] Loss: 0.000005 Train Epoch: 4 [14080/104000 (14%)] Loss: 0.000019 Train Epoch: 4 [14720/104000 (14%)] Loss: 0.000234 Train Epoch: 4 [15360/104000 (15%)] Loss: 0.000008 Train Epoch: 4 [16000/104000 (15%)] Loss: 0.000007 Train Epoch: 4 [16640/104000 (16%)] Loss: 0.000005 Train Epoch: 4 [17280/104000 (17%)] Loss: 0.000007 Train Epoch: 4 [17920/104000 (17%)] Loss: 0.000048 Train Epoch: 4 [18560/104000 (18%)] Loss: 0.000016 Train Epoch: 4 [19200/104000 (18%)] Loss: 0.000005 Train Epoch: 4 [19840/104000 (19%)] Loss: 0.000004 Train Epoch: 4 [20480/104000 (20%)] Loss: 0.000050 Train Epoch: 4 [21120/104000 (20%)] Loss: 0.000004 Train Epoch: 4 [21760/104000 (21%)] Loss: 0.000017 Train Epoch: 4 [22400/104000 (22%)] Loss: 0.000107 Train Epoch: 4 [23040/104000 (22%)] Loss: 0.000007 Train Epoch: 4 [23680/104000 (23%)] Loss: 0.000007 Train Epoch: 4 [24320/104000 (23%)] Loss: 0.000073 Train Epoch: 4 [24960/104000 (24%)] Loss: 0.000004 Train Epoch: 4 [25600/104000 (25%)] Loss: 0.000052 Train Epoch: 4 [26240/104000 (25%)] Loss: 0.000008 Train Epoch: 4 [26880/104000 (26%)] Loss: 0.000021 Train Epoch: 4 [27520/104000 (26%)] Loss: 0.000006 Train Epoch: 4 [28160/104000 (27%)] Loss: 0.000038 Train Epoch: 4 [28800/104000 (28%)] Loss: 0.000030 Train Epoch: 4 [29440/104000 (28%)] Loss: 0.000009 Train Epoch: 4 [30080/104000 (29%)] Loss: 0.000024 Train Epoch: 4 [30720/104000 (30%)] Loss: 0.000007 Train Epoch: 4 [31360/104000 (30%)] Loss: 0.000018 Train Epoch: 4 [32000/104000 (31%)] Loss: 0.000010 Train Epoch: 4 [32640/104000 (31%)] Loss: 0.000007 Train Epoch: 4 [33280/104000 (32%)] Loss: 0.000004 Train Epoch: 4 [33920/104000 (33%)] Loss: 0.000004 Train Epoch: 4 [34560/104000 (33%)] Loss: 0.000003 Train Epoch: 4 [35200/104000 (34%)] Loss: 0.000005 Train Epoch: 4 [35840/104000 (34%)] Loss: 0.000014 Train Epoch: 4 [36480/104000 (35%)] Loss: 0.000164 Train Epoch: 4 [37120/104000 (36%)] Loss: 0.000014 Train Epoch: 4 [37760/104000 (36%)] Loss: 0.000007 Train Epoch: 4 [38400/104000 (37%)] Loss: 0.000030 Train Epoch: 4 [39040/104000 (38%)] Loss: 0.000005 Train Epoch: 4 [39680/104000 (38%)] Loss: 0.000012 Train Epoch: 4 [40320/104000 (39%)] Loss: 0.000009 Train Epoch: 4 [40960/104000 (39%)] Loss: 0.000049 Train Epoch: 4 [41600/104000 (40%)] Loss: 0.000006 Train Epoch: 4 [42240/104000 (41%)] Loss: 0.000011 Train Epoch: 4 [42880/104000 (41%)] Loss: 0.000004 Train Epoch: 4 [43520/104000 (42%)] Loss: 0.000008 Train Epoch: 4 [44160/104000 (42%)] Loss: 0.000014 Train Epoch: 4 [44800/104000 (43%)] Loss: 0.000010 Train Epoch: 4 [45440/104000 (44%)] Loss: 0.000005 Train Epoch: 4 [46080/104000 (44%)] Loss: 0.000009 Train Epoch: 4 [46720/104000 (45%)] Loss: 0.000004 Train Epoch: 4 [47360/104000 (46%)] Loss: 0.000004 Train Epoch: 4 [48000/104000 (46%)] Loss: 0.000007 Train Epoch: 4 [48640/104000 (47%)] Loss: 0.000008 Train Epoch: 4 [49280/104000 (47%)] Loss: 0.000095 Train Epoch: 4 [49920/104000 (48%)] Loss: 0.000013 Train Epoch: 4 [50560/104000 (49%)] Loss: 0.000014 Train Epoch: 4 [51200/104000 (49%)] Loss: 0.000009 Train Epoch: 4 [51840/104000 (50%)] Loss: 0.000049 Train Epoch: 4 [52480/104000 (50%)] Loss: 0.000016 Train Epoch: 4 [53120/104000 (51%)] Loss: 0.000009 Train Epoch: 4 [53760/104000 (52%)] Loss: 0.000007 Train Epoch: 4 [54400/104000 (52%)] Loss: 0.000007 Train Epoch: 4 [55040/104000 (53%)] Loss: 0.000116 Train Epoch: 4 [55680/104000 (54%)] Loss: 0.000004 Train Epoch: 4 [56320/104000 (54%)] Loss: 0.000007 Train Epoch: 4 [56960/104000 (55%)] Loss: 0.000013 Train Epoch: 4 [57600/104000 (55%)] Loss: 0.000011 Train Epoch: 4 [58240/104000 (56%)] Loss: 0.000013 Train Epoch: 4 [58880/104000 (57%)] Loss: 0.000005 Train Epoch: 4 [59520/104000 (57%)] Loss: 0.000028 Train Epoch: 4 [60160/104000 (58%)] Loss: 0.000005 Train Epoch: 4 [60800/104000 (58%)] Loss: 0.000034 Train Epoch: 4 [61440/104000 (59%)] Loss: 0.000006 Train Epoch: 4 [62080/104000 (60%)] Loss: 0.000023 Train Epoch: 4 [62720/104000 (60%)] Loss: 0.000006 Train Epoch: 4 [63360/104000 (61%)] Loss: 0.000006 Train Epoch: 4 [64000/104000 (62%)] Loss: 0.000006 Train Epoch: 4 [64640/104000 (62%)] Loss: 0.000011 Train Epoch: 4 [65280/104000 (63%)] Loss: 0.000011 Train Epoch: 4 [65920/104000 (63%)] Loss: 0.000006 Train Epoch: 4 [66560/104000 (64%)] Loss: 0.000030 Train Epoch: 4 [67200/104000 (65%)] Loss: 0.000004 Train Epoch: 4 [67840/104000 (65%)] Loss: 0.000005 Train Epoch: 4 [68480/104000 (66%)] Loss: 0.000010 Train Epoch: 4 [69120/104000 (66%)] Loss: 0.000015 Train Epoch: 4 [69760/104000 (67%)] Loss: 0.000004 Train Epoch: 4 [70400/104000 (68%)] Loss: 0.000012 Train Epoch: 4 [71040/104000 (68%)] Loss: 0.000024 Train Epoch: 4 [71680/104000 (69%)] Loss: 0.000081 Train Epoch: 4 [72320/104000 (70%)] Loss: 0.000004 Train Epoch: 4 [72960/104000 (70%)] Loss: 0.000009 Train Epoch: 4 [73600/104000 (71%)] Loss: 0.000017 Train Epoch: 4 [74240/104000 (71%)] Loss: 0.000035 Train Epoch: 4 [74880/104000 (72%)] Loss: 0.000089 Train Epoch: 4 [75520/104000 (73%)] Loss: 0.000005 Train Epoch: 4 [76160/104000 (73%)] Loss: 0.000008 Train Epoch: 4 [76800/104000 (74%)] Loss: 0.000008 Train Epoch: 4 [77440/104000 (74%)] Loss: 0.000010 Train Epoch: 4 [78080/104000 (75%)] Loss: 0.000013 Train Epoch: 4 [78720/104000 (76%)] Loss: 0.000014 Train Epoch: 4 [79360/104000 (76%)] Loss: 0.000064 Train Epoch: 4 [80000/104000 (77%)] Loss: 0.000233 Train Epoch: 4 [80640/104000 (78%)] Loss: 0.000006 Train Epoch: 4 [81280/104000 (78%)] Loss: 0.000103 Train Epoch: 4 [81920/104000 (79%)] Loss: 0.000003 Train Epoch: 4 [82560/104000 (79%)] Loss: 0.000009 Train Epoch: 4 [83200/104000 (80%)] Loss: 0.000061 Train Epoch: 4 [83840/104000 (81%)] Loss: 0.000067 Train Epoch: 4 [84480/104000 (81%)] Loss: 0.000004 Train Epoch: 4 [85120/104000 (82%)] Loss: 0.000003 Train Epoch: 4 [85760/104000 (82%)] Loss: 0.000018 Train Epoch: 4 [86400/104000 (83%)] Loss: 0.000004 Train Epoch: 4 [87040/104000 (84%)] Loss: 0.000019 Train Epoch: 4 [87680/104000 (84%)] Loss: 0.000004 Train Epoch: 4 [88320/104000 (85%)] Loss: 0.000081 Train Epoch: 4 [88960/104000 (86%)] Loss: 0.000031 Train Epoch: 4 [89600/104000 (86%)] Loss: 0.000005 Train Epoch: 4 [90240/104000 (87%)] Loss: 0.000009 Train Epoch: 4 [90880/104000 (87%)] Loss: 0.000009 Train Epoch: 4 [91520/104000 (88%)] Loss: 0.000006 Train Epoch: 4 [92160/104000 (89%)] Loss: 0.000010 Train Epoch: 4 [92800/104000 (89%)] Loss: 0.000007 Train Epoch: 4 [93440/104000 (90%)] Loss: 0.000014 Train Epoch: 4 [94080/104000 (90%)] Loss: 0.000004 Train Epoch: 4 [94720/104000 (91%)] Loss: 0.000014 Train Epoch: 4 [95360/104000 (92%)] Loss: 0.000002 Train Epoch: 4 [96000/104000 (92%)] Loss: 0.000009 Train Epoch: 4 [96640/104000 (93%)] Loss: 0.000005 Train Epoch: 4 [97280/104000 (94%)] Loss: 0.000002 Train Epoch: 4 [97920/104000 (94%)] Loss: 0.000010 Train Epoch: 4 [98560/104000 (95%)] Loss: 0.000021 Train Epoch: 4 [99200/104000 (95%)] Loss: 0.000141 Train Epoch: 4 [99840/104000 (96%)] Loss: 0.000010 Train Epoch: 4 [100480/104000 (97%)] Loss: 0.000076 Train Epoch: 4 [101120/104000 (97%)] Loss: 0.000013 Train Epoch: 4 [101760/104000 (98%)] Loss: 0.000016 Train Epoch: 4 [102400/104000 (98%)] Loss: 0.000010 Train Epoch: 4 [103040/104000 (99%)] Loss: 0.000018 Train Epoch: 4 [103680/104000 (100%)] Loss: 0.000008 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.3767, Accuracy: 24775/30000 (83%) Train Epoch: 5 [0/104000 (0%)] Loss: 0.000031 Train Epoch: 5 [640/104000 (1%)] Loss: 0.000005 Train Epoch: 5 [1280/104000 (1%)] Loss: 0.000005 Train Epoch: 5 [1920/104000 (2%)] Loss: 0.000002 Train Epoch: 5 [2560/104000 (2%)] Loss: 0.000023 Train Epoch: 5 [3200/104000 (3%)] Loss: 0.000010 Train Epoch: 5 [3840/104000 (4%)] Loss: 0.000008 Train Epoch: 5 [4480/104000 (4%)] Loss: 0.000003 Train Epoch: 5 [5120/104000 (5%)] Loss: 0.000037 Train Epoch: 5 [5760/104000 (6%)] Loss: 0.000203 Train Epoch: 5 [6400/104000 (6%)] Loss: 0.000017 Train Epoch: 5 [7040/104000 (7%)] Loss: 0.000013 Train Epoch: 5 [7680/104000 (7%)] Loss: 0.000004 Train Epoch: 5 [8320/104000 (8%)] Loss: 0.000015 Train Epoch: 5 [8960/104000 (9%)] Loss: 0.000015 Train Epoch: 5 [9600/104000 (9%)] Loss: 0.000012 Train Epoch: 5 [10240/104000 (10%)] Loss: 0.000010 Train Epoch: 5 [10880/104000 (10%)] Loss: 0.000012 Train Epoch: 5 [11520/104000 (11%)] Loss: 0.000009 Train Epoch: 5 [12160/104000 (12%)] Loss: 0.000044 Train Epoch: 5 [12800/104000 (12%)] Loss: 0.000008 Train Epoch: 5 [13440/104000 (13%)] Loss: 0.000007 Train Epoch: 5 [14080/104000 (14%)] Loss: 0.000004 Train Epoch: 5 [14720/104000 (14%)] Loss: 0.000002 Train Epoch: 5 [15360/104000 (15%)] Loss: 0.000009 Train Epoch: 5 [16000/104000 (15%)] Loss: 0.000021 Train Epoch: 5 [16640/104000 (16%)] Loss: 0.000008 Train Epoch: 5 [17280/104000 (17%)] Loss: 0.000050 Train Epoch: 5 [17920/104000 (17%)] Loss: 0.000018 Train Epoch: 5 [18560/104000 (18%)] Loss: 0.000006 Train Epoch: 5 [19200/104000 (18%)] Loss: 0.000005 Train Epoch: 5 [19840/104000 (19%)] Loss: 0.000004 Train Epoch: 5 [20480/104000 (20%)] Loss: 0.000008 Train Epoch: 5 [21120/104000 (20%)] Loss: 0.000010 Train Epoch: 5 [21760/104000 (21%)] Loss: 0.000006 Train Epoch: 5 [22400/104000 (22%)] Loss: 0.000014 Train Epoch: 5 [23040/104000 (22%)] Loss: 0.000005 Train Epoch: 5 [23680/104000 (23%)] Loss: 0.000004 Train Epoch: 5 [24320/104000 (23%)] Loss: 0.000003 Train Epoch: 5 [24960/104000 (24%)] Loss: 0.000056 Train Epoch: 5 [25600/104000 (25%)] Loss: 0.000129 Train Epoch: 5 [26240/104000 (25%)] Loss: 0.000009 Train Epoch: 5 [26880/104000 (26%)] Loss: 0.000006 Train Epoch: 5 [27520/104000 (26%)] Loss: 0.000022 Train Epoch: 5 [28160/104000 (27%)] Loss: 0.000006 Train Epoch: 5 [28800/104000 (28%)] Loss: 0.000012 Train Epoch: 5 [29440/104000 (28%)] Loss: 0.000008 Train Epoch: 5 [30080/104000 (29%)] Loss: 0.000037 Train Epoch: 5 [30720/104000 (30%)] Loss: 0.000003 Train Epoch: 5 [31360/104000 (30%)] Loss: 0.000004 Train Epoch: 5 [32000/104000 (31%)] Loss: 0.000018 Train Epoch: 5 [32640/104000 (31%)] Loss: 0.000014 Train Epoch: 5 [33280/104000 (32%)] Loss: 0.000010 Train Epoch: 5 [33920/104000 (33%)] Loss: 0.000005 Train Epoch: 5 [34560/104000 (33%)] Loss: 0.000004 Train Epoch: 5 [35200/104000 (34%)] Loss: 0.000007 Train Epoch: 5 [35840/104000 (34%)] Loss: 0.000004 Train Epoch: 5 [36480/104000 (35%)] Loss: 0.000015 Train Epoch: 5 [37120/104000 (36%)] Loss: 0.000008 Train Epoch: 5 [37760/104000 (36%)] Loss: 0.000004 Train Epoch: 5 [38400/104000 (37%)] Loss: 0.000010 Train Epoch: 5 [39040/104000 (38%)] Loss: 0.000002 Train Epoch: 5 [39680/104000 (38%)] Loss: 0.000051 Train Epoch: 5 [40320/104000 (39%)] Loss: 0.000012 Train Epoch: 5 [40960/104000 (39%)] Loss: 0.000006 Train Epoch: 5 [41600/104000 (40%)] Loss: 0.000031 Train Epoch: 5 [42240/104000 (41%)] Loss: 0.000008 Train Epoch: 5 [42880/104000 (41%)] Loss: 0.000015 Train Epoch: 5 [43520/104000 (42%)] Loss: 0.000006 Train Epoch: 5 [44160/104000 (42%)] Loss: 0.000035 Train Epoch: 5 [44800/104000 (43%)] Loss: 0.000011 Train Epoch: 5 [45440/104000 (44%)] Loss: 0.000012 Train Epoch: 5 [46080/104000 (44%)] Loss: 0.000006 Train Epoch: 5 [46720/104000 (45%)] Loss: 0.000009 Train Epoch: 5 [47360/104000 (46%)] Loss: 0.000037 Train Epoch: 5 [48000/104000 (46%)] Loss: 0.000002 Train Epoch: 5 [48640/104000 (47%)] Loss: 0.000005 Train Epoch: 5 [49280/104000 (47%)] Loss: 0.000002 Train Epoch: 5 [49920/104000 (48%)] Loss: 0.000012 Train Epoch: 5 [50560/104000 (49%)] Loss: 0.000006 Train Epoch: 5 [51200/104000 (49%)] Loss: 0.000061 Train Epoch: 5 [51840/104000 (50%)] Loss: 0.000006 Train Epoch: 5 [52480/104000 (50%)] Loss: 0.000004 Train Epoch: 5 [53120/104000 (51%)] Loss: 0.000059 Train Epoch: 5 [53760/104000 (52%)] Loss: 0.000010 Train Epoch: 5 [54400/104000 (52%)] Loss: 0.000003 Train Epoch: 5 [55040/104000 (53%)] Loss: 0.000016 Train Epoch: 5 [55680/104000 (54%)] Loss: 0.000005 Train Epoch: 5 [56320/104000 (54%)] Loss: 0.000006 Train Epoch: 5 [56960/104000 (55%)] Loss: 0.000227 Train Epoch: 5 [57600/104000 (55%)] Loss: 0.000008 Train Epoch: 5 [58240/104000 (56%)] Loss: 0.000008 Train Epoch: 5 [58880/104000 (57%)] Loss: 0.000014 Train Epoch: 5 [59520/104000 (57%)] Loss: 0.000002 Train Epoch: 5 [60160/104000 (58%)] Loss: 0.000005 Train Epoch: 5 [60800/104000 (58%)] Loss: 0.000031 Train Epoch: 5 [61440/104000 (59%)] Loss: 0.000004 Train Epoch: 5 [62080/104000 (60%)] Loss: 0.000013 Train Epoch: 5 [62720/104000 (60%)] Loss: 0.000020 Train Epoch: 5 [63360/104000 (61%)] Loss: 0.000035 Train Epoch: 5 [64000/104000 (62%)] Loss: 0.000007 Train Epoch: 5 [64640/104000 (62%)] Loss: 0.000008 Train Epoch: 5 [65280/104000 (63%)] Loss: 0.000005 Train Epoch: 5 [65920/104000 (63%)] Loss: 0.000034 Train Epoch: 5 [66560/104000 (64%)] Loss: 0.000004 Train Epoch: 5 [67200/104000 (65%)] Loss: 0.000054 Train Epoch: 5 [67840/104000 (65%)] Loss: 0.000040 Train Epoch: 5 [68480/104000 (66%)] Loss: 0.000004 Train Epoch: 5 [69120/104000 (66%)] Loss: 0.000003 Train Epoch: 5 [69760/104000 (67%)] Loss: 0.000023 Train Epoch: 5 [70400/104000 (68%)] Loss: 0.000005 Train Epoch: 5 [71040/104000 (68%)] Loss: 0.000037 Train Epoch: 5 [71680/104000 (69%)] Loss: 0.000005 Train Epoch: 5 [72320/104000 (70%)] Loss: 0.000007 Train Epoch: 5 [72960/104000 (70%)] Loss: 0.000005 Train Epoch: 5 [73600/104000 (71%)] Loss: 0.000004 Train Epoch: 5 [74240/104000 (71%)] Loss: 0.000008 Train Epoch: 5 [74880/104000 (72%)] Loss: 0.000015 Train Epoch: 5 [75520/104000 (73%)] Loss: 0.000004 Train Epoch: 5 [76160/104000 (73%)] Loss: 0.000005 Train Epoch: 5 [76800/104000 (74%)] Loss: 0.000004 Train Epoch: 5 [77440/104000 (74%)] Loss: 0.000003 Train Epoch: 5 [78080/104000 (75%)] Loss: 0.000013 Train Epoch: 5 [78720/104000 (76%)] Loss: 0.000002 Train Epoch: 5 [79360/104000 (76%)] Loss: 0.000189 Train Epoch: 5 [80000/104000 (77%)] Loss: 0.000044 Train Epoch: 5 [80640/104000 (78%)] Loss: 0.000009 Train Epoch: 5 [81280/104000 (78%)] Loss: 0.000002 Train Epoch: 5 [81920/104000 (79%)] Loss: 0.000008 Train Epoch: 5 [82560/104000 (79%)] Loss: 0.000002 Train Epoch: 5 [83200/104000 (80%)] Loss: 0.000004 Train Epoch: 5 [83840/104000 (81%)] Loss: 0.000019 Train Epoch: 5 [84480/104000 (81%)] Loss: 0.000004 Train Epoch: 5 [85120/104000 (82%)] Loss: 0.000004 Train Epoch: 5 [85760/104000 (82%)] Loss: 0.000009 Train Epoch: 5 [86400/104000 (83%)] Loss: 0.000012 Train Epoch: 5 [87040/104000 (84%)] Loss: 0.000016 Train Epoch: 5 [87680/104000 (84%)] Loss: 0.000008 Train Epoch: 5 [88320/104000 (85%)] Loss: 0.000025 Train Epoch: 5 [88960/104000 (86%)] Loss: 0.000006 Train Epoch: 5 [89600/104000 (86%)] Loss: 0.000398 Train Epoch: 5 [90240/104000 (87%)] Loss: 0.000003 Train Epoch: 5 [90880/104000 (87%)] Loss: 0.000003 Train Epoch: 5 [91520/104000 (88%)] Loss: 0.000008 Train Epoch: 5 [92160/104000 (89%)] Loss: 0.000012 Train Epoch: 5 [92800/104000 (89%)] Loss: 0.000009 Train Epoch: 5 [93440/104000 (90%)] Loss: 0.000018 Train Epoch: 5 [94080/104000 (90%)] Loss: 0.000004 Train Epoch: 5 [94720/104000 (91%)] Loss: 0.000007 Train Epoch: 5 [95360/104000 (92%)] Loss: 0.000002 Train Epoch: 5 [96000/104000 (92%)] Loss: 0.000007 Train Epoch: 5 [96640/104000 (93%)] Loss: 0.000005 Train Epoch: 5 [97280/104000 (94%)] Loss: 0.000012 Train Epoch: 5 [97920/104000 (94%)] Loss: 0.000004 Train Epoch: 5 [98560/104000 (95%)] Loss: 0.000007 Train Epoch: 5 [99200/104000 (95%)] Loss: 0.000391 Train Epoch: 5 [99840/104000 (96%)] Loss: 0.000004 Train Epoch: 5 [100480/104000 (97%)] Loss: 0.000006 Train Epoch: 5 [101120/104000 (97%)] Loss: 0.000467 Train Epoch: 5 [101760/104000 (98%)] Loss: 0.000007 Train Epoch: 5 [102400/104000 (98%)] Loss: 0.000004 Train Epoch: 5 [103040/104000 (99%)] Loss: 0.000067 Train Epoch: 5 [103680/104000 (100%)] Loss: 0.000057 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.3028, Accuracy: 25782/30000 (86%) [100.0, 100.0, 100.0, 100.0] [88.56, 91.44666666666667, 90.16666666666667, 85.94] The number of neurons in CNN layer is 50 Train Epoch: 1 [0/104000 (0%)] Loss: 1.178634 Train Epoch: 1 [640/104000 (1%)] Loss: 0.000044 Train Epoch: 1 [1280/104000 (1%)] Loss: 0.000000 Train Epoch: 1 [1920/104000 (2%)] Loss: 0.000000 Train Epoch: 1 [2560/104000 (2%)] Loss: 0.000002 Train Epoch: 1 [3200/104000 (3%)] Loss: 0.000001 Train Epoch: 1 [3840/104000 (4%)] Loss: 0.000000 Train Epoch: 1 [4480/104000 (4%)] Loss: 0.000000 Train Epoch: 1 [5120/104000 (5%)] Loss: 0.000000 Train Epoch: 1 [5760/104000 (6%)] Loss: 0.000000 Train Epoch: 1 [6400/104000 (6%)] Loss: 0.000000 Train Epoch: 1 [7040/104000 (7%)] Loss: 0.000010 Train Epoch: 1 [7680/104000 (7%)] Loss: 0.000000 Train Epoch: 1 [8320/104000 (8%)] Loss: 0.000000 Train Epoch: 1 [8960/104000 (9%)] Loss: 0.000000 Train Epoch: 1 [9600/104000 (9%)] Loss: 0.000001 Train Epoch: 1 [10240/104000 (10%)] Loss: 0.000000 Train Epoch: 1 [10880/104000 (10%)] Loss: 0.000000 Train Epoch: 1 [11520/104000 (11%)] Loss: 0.000000 Train Epoch: 1 [12160/104000 (12%)] Loss: 0.000000 Train Epoch: 1 [12800/104000 (12%)] Loss: 0.000000 Train Epoch: 1 [13440/104000 (13%)] Loss: 0.000002 Train Epoch: 1 [14080/104000 (14%)] Loss: 0.000000 Train Epoch: 1 [14720/104000 (14%)] Loss: 0.000000 Train Epoch: 1 [15360/104000 (15%)] Loss: 0.000000 Train Epoch: 1 [16000/104000 (15%)] Loss: 0.000000 Train Epoch: 1 [16640/104000 (16%)] Loss: 0.000000 Train Epoch: 1 [17280/104000 (17%)] Loss: 0.000000 Train Epoch: 1 [17920/104000 (17%)] Loss: 0.000000 Train Epoch: 1 [18560/104000 (18%)] Loss: 0.000000 Train Epoch: 1 [19200/104000 (18%)] Loss: 0.000000 Train Epoch: 1 [19840/104000 (19%)] Loss: 0.000000 Train Epoch: 1 [20480/104000 (20%)] Loss: 0.000000 Train Epoch: 1 [21120/104000 (20%)] Loss: 0.000000 Train Epoch: 1 [21760/104000 (21%)] Loss: 0.000000 Train Epoch: 1 [22400/104000 (22%)] Loss: 0.000000 Train Epoch: 1 [23040/104000 (22%)] Loss: 0.000000 Train Epoch: 1 [23680/104000 (23%)] Loss: 0.000000 Train Epoch: 1 [24320/104000 (23%)] Loss: 0.000000 Train Epoch: 1 [24960/104000 (24%)] Loss: 0.000003 Train Epoch: 1 [25600/104000 (25%)] Loss: 0.000000 Train Epoch: 1 [26240/104000 (25%)] Loss: 0.000000 Train Epoch: 1 [26880/104000 (26%)] Loss: 0.000000 Train Epoch: 1 [27520/104000 (26%)] Loss: 0.000000 Train Epoch: 1 [28160/104000 (27%)] Loss: 0.000004 Train Epoch: 1 [28800/104000 (28%)] Loss: 0.000010 Train Epoch: 1 [29440/104000 (28%)] Loss: 0.000000 Train Epoch: 1 [30080/104000 (29%)] Loss: 0.000000 Train Epoch: 1 [30720/104000 (30%)] Loss: 0.000000 Train Epoch: 1 [31360/104000 (30%)] Loss: 0.000000 Train Epoch: 1 [32000/104000 (31%)] Loss: 0.000000 Train Epoch: 1 [32640/104000 (31%)] Loss: 0.000000 Train Epoch: 1 [33280/104000 (32%)] Loss: 0.000000 Train Epoch: 1 [33920/104000 (33%)] Loss: 0.000000 Train Epoch: 1 [34560/104000 (33%)] Loss: 0.000000 Train Epoch: 1 [35200/104000 (34%)] Loss: 0.000000 Train Epoch: 1 [35840/104000 (34%)] Loss: 0.000000 Train Epoch: 1 [36480/104000 (35%)] Loss: 0.000000 Train Epoch: 1 [37120/104000 (36%)] Loss: 0.000000 Train Epoch: 1 [37760/104000 (36%)] Loss: 0.000000 Train Epoch: 1 [38400/104000 (37%)] Loss: 0.000000 Train Epoch: 1 [39040/104000 (38%)] Loss: 0.000000 Train Epoch: 1 [39680/104000 (38%)] Loss: 0.000000 Train Epoch: 1 [40320/104000 (39%)] Loss: 0.000003 Train Epoch: 1 [40960/104000 (39%)] Loss: 0.000000 Train Epoch: 1 [41600/104000 (40%)] Loss: 0.000000 Train Epoch: 1 [42240/104000 (41%)] Loss: 0.000000 Train Epoch: 1 [42880/104000 (41%)] Loss: 0.000000 Train Epoch: 1 [43520/104000 (42%)] Loss: 0.000000 Train Epoch: 1 [44160/104000 (42%)] Loss: 0.000000 Train Epoch: 1 [44800/104000 (43%)] Loss: 0.000000 Train Epoch: 1 [45440/104000 (44%)] Loss: 0.000001 Train Epoch: 1 [46080/104000 (44%)] Loss: 0.000000 Train Epoch: 1 [46720/104000 (45%)] Loss: 0.000000 Train Epoch: 1 [47360/104000 (46%)] Loss: 0.000000 Train Epoch: 1 [48000/104000 (46%)] Loss: 0.000000 Train Epoch: 1 [48640/104000 (47%)] Loss: 0.000003 Train Epoch: 1 [49280/104000 (47%)] Loss: 0.000000 Train Epoch: 1 [49920/104000 (48%)] Loss: 0.000000 Train Epoch: 1 [50560/104000 (49%)] Loss: 0.000000 Train Epoch: 1 [51200/104000 (49%)] Loss: 0.000000 Train Epoch: 1 [51840/104000 (50%)] Loss: 0.000000 Train Epoch: 1 [52480/104000 (50%)] Loss: 0.000000 Train Epoch: 1 [53120/104000 (51%)] Loss: 0.000000 Train Epoch: 1 [53760/104000 (52%)] Loss: 0.000000 Train Epoch: 1 [54400/104000 (52%)] Loss: 0.000000 Train Epoch: 1 [55040/104000 (53%)] Loss: 0.000001 Train Epoch: 1 [55680/104000 (54%)] Loss: 0.000000 Train Epoch: 1 [56320/104000 (54%)] Loss: 0.000000 Train Epoch: 1 [56960/104000 (55%)] Loss: 0.000000 Train Epoch: 1 [57600/104000 (55%)] Loss: 0.000000 Train Epoch: 1 [58240/104000 (56%)] Loss: 0.000000 Train Epoch: 1 [58880/104000 (57%)] Loss: 0.000000 Train Epoch: 1 [59520/104000 (57%)] Loss: 0.000000 Train Epoch: 1 [60160/104000 (58%)] Loss: 0.000000 Train Epoch: 1 [60800/104000 (58%)] Loss: 0.000000 Train Epoch: 1 [61440/104000 (59%)] Loss: 0.000000 Train Epoch: 1 [62080/104000 (60%)] Loss: 0.000000 Train Epoch: 1 [62720/104000 (60%)] Loss: 0.000000 Train Epoch: 1 [63360/104000 (61%)] Loss: 0.000000 Train Epoch: 1 [64000/104000 (62%)] Loss: 0.000000 Train Epoch: 1 [64640/104000 (62%)] Loss: 0.000000 Train Epoch: 1 [65280/104000 (63%)] Loss: 0.000000 Train Epoch: 1 [65920/104000 (63%)] Loss: 0.000000 Train Epoch: 1 [66560/104000 (64%)] Loss: 0.000000 Train Epoch: 1 [67200/104000 (65%)] Loss: 0.000000 Train Epoch: 1 [67840/104000 (65%)] Loss: 0.000000 Train Epoch: 1 [68480/104000 (66%)] Loss: 0.000000 Train Epoch: 1 [69120/104000 (66%)] Loss: 0.000000 Train Epoch: 1 [69760/104000 (67%)] Loss: 0.000000 Train Epoch: 1 [70400/104000 (68%)] Loss: 0.000000 Train Epoch: 1 [71040/104000 (68%)] Loss: 0.000000 Train Epoch: 1 [71680/104000 (69%)] Loss: 0.000000 Train Epoch: 1 [72320/104000 (70%)] Loss: 0.000000 Train Epoch: 1 [72960/104000 (70%)] Loss: 0.000000 Train Epoch: 1 [73600/104000 (71%)] Loss: 0.000003 Train Epoch: 1 [74240/104000 (71%)] Loss: 0.000000 Train Epoch: 1 [74880/104000 (72%)] Loss: 0.000000 Train Epoch: 1 [75520/104000 (73%)] Loss: 0.000000 Train Epoch: 1 [76160/104000 (73%)] Loss: 0.000002 Train Epoch: 1 [76800/104000 (74%)] Loss: 0.000000 Train Epoch: 1 [77440/104000 (74%)] Loss: 0.000000 Train Epoch: 1 [78080/104000 (75%)] Loss: 0.000000 Train Epoch: 1 [78720/104000 (76%)] Loss: 0.000000 Train Epoch: 1 [79360/104000 (76%)] Loss: 0.000000 Train Epoch: 1 [80000/104000 (77%)] Loss: 0.000000 Train Epoch: 1 [80640/104000 (78%)] Loss: 0.000000 Train Epoch: 1 [81280/104000 (78%)] Loss: 0.000001 Train Epoch: 1 [81920/104000 (79%)] Loss: 0.000000 Train Epoch: 1 [82560/104000 (79%)] Loss: 0.000000 Train Epoch: 1 [83200/104000 (80%)] Loss: 0.000000 Train Epoch: 1 [83840/104000 (81%)] Loss: 0.000000 Train Epoch: 1 [84480/104000 (81%)] Loss: 0.000000 Train Epoch: 1 [85120/104000 (82%)] Loss: 0.000000 Train Epoch: 1 [85760/104000 (82%)] Loss: 0.000000 Train Epoch: 1 [86400/104000 (83%)] Loss: 0.000000 Train Epoch: 1 [87040/104000 (84%)] Loss: 0.000000 Train Epoch: 1 [87680/104000 (84%)] Loss: 0.000000 Train Epoch: 1 [88320/104000 (85%)] Loss: 0.000000 Train Epoch: 1 [88960/104000 (86%)] Loss: 0.000000 Train Epoch: 1 [89600/104000 (86%)] Loss: 0.000000 Train Epoch: 1 [90240/104000 (87%)] Loss: 0.000000 Train Epoch: 1 [90880/104000 (87%)] Loss: 0.000000 Train Epoch: 1 [91520/104000 (88%)] Loss: 0.000001 Train Epoch: 1 [92160/104000 (89%)] Loss: 0.000000 Train Epoch: 1 [92800/104000 (89%)] Loss: 0.000000 Train Epoch: 1 [93440/104000 (90%)] Loss: 0.000000 Train Epoch: 1 [94080/104000 (90%)] Loss: 0.000000 Train Epoch: 1 [94720/104000 (91%)] Loss: 0.000000 Train Epoch: 1 [95360/104000 (92%)] Loss: 0.000000 Train Epoch: 1 [96000/104000 (92%)] Loss: 0.000000 Train Epoch: 1 [96640/104000 (93%)] Loss: 0.000000 Train Epoch: 1 [97280/104000 (94%)] Loss: 0.000000 Train Epoch: 1 [97920/104000 (94%)] Loss: 0.000000 Train Epoch: 1 [98560/104000 (95%)] Loss: 0.000000 Train Epoch: 1 [99200/104000 (95%)] Loss: 0.000000 Train Epoch: 1 [99840/104000 (96%)] Loss: 0.000000 Train Epoch: 1 [100480/104000 (97%)] Loss: 0.000000 Train Epoch: 1 [101120/104000 (97%)] Loss: 0.000000 Train Epoch: 1 [101760/104000 (98%)] Loss: 0.000000 Train Epoch: 1 [102400/104000 (98%)] Loss: 0.000000 Train Epoch: 1 [103040/104000 (99%)] Loss: 0.000000 Train Epoch: 1 [103680/104000 (100%)] Loss: 0.000000 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 1.3950, Accuracy: 24668/30000 (82%) Train Epoch: 2 [0/104000 (0%)] Loss: 0.000000 Train Epoch: 2 [640/104000 (1%)] Loss: 0.000000 Train Epoch: 2 [1280/104000 (1%)] Loss: 0.000000 Train Epoch: 2 [1920/104000 (2%)] Loss: 0.000000 Train Epoch: 2 [2560/104000 (2%)] Loss: 0.000000 Train Epoch: 2 [3200/104000 (3%)] Loss: 0.000000 Train Epoch: 2 [3840/104000 (4%)] Loss: 0.000000 Train Epoch: 2 [4480/104000 (4%)] Loss: 0.000000 Train Epoch: 2 [5120/104000 (5%)] Loss: 0.000000 Train Epoch: 2 [5760/104000 (6%)] Loss: 0.000000 Train Epoch: 2 [6400/104000 (6%)] Loss: 0.000000 Train Epoch: 2 [7040/104000 (7%)] Loss: 0.000000 Train Epoch: 2 [7680/104000 (7%)] Loss: 0.000000 Train Epoch: 2 [8320/104000 (8%)] Loss: 0.000000 Train Epoch: 2 [8960/104000 (9%)] Loss: 0.000000 Train Epoch: 2 [9600/104000 (9%)] Loss: 0.000000 Train Epoch: 2 [10240/104000 (10%)] Loss: 0.000000 Train Epoch: 2 [10880/104000 (10%)] Loss: 0.000000 Train Epoch: 2 [11520/104000 (11%)] Loss: 0.000000 Train Epoch: 2 [12160/104000 (12%)] Loss: 0.000000 Train Epoch: 2 [12800/104000 (12%)] Loss: 0.000000 Train Epoch: 2 [13440/104000 (13%)] Loss: 0.000000 Train Epoch: 2 [14080/104000 (14%)] Loss: 0.000000 Train Epoch: 2 [14720/104000 (14%)] Loss: 0.000000 Train Epoch: 2 [15360/104000 (15%)] Loss: 0.000000 Train Epoch: 2 [16000/104000 (15%)] Loss: 0.000000 Train Epoch: 2 [16640/104000 (16%)] Loss: 0.000000 Train Epoch: 2 [17280/104000 (17%)] Loss: 0.000000 Train Epoch: 2 [17920/104000 (17%)] Loss: 0.000000 Train Epoch: 2 [18560/104000 (18%)] Loss: 0.000000 Train Epoch: 2 [19200/104000 (18%)] Loss: 0.000000 Train Epoch: 2 [19840/104000 (19%)] Loss: 0.000000 Train Epoch: 2 [20480/104000 (20%)] Loss: 0.000000 Train Epoch: 2 [21120/104000 (20%)] Loss: 0.000000 Train Epoch: 2 [21760/104000 (21%)] Loss: 0.000000 Train Epoch: 2 [22400/104000 (22%)] Loss: 0.000000 Train Epoch: 2 [23040/104000 (22%)] Loss: 0.000000 Train Epoch: 2 [23680/104000 (23%)] Loss: 0.000000 Train Epoch: 2 [24320/104000 (23%)] Loss: 0.000000 Train Epoch: 2 [24960/104000 (24%)] Loss: 0.000000 Train Epoch: 2 [25600/104000 (25%)] Loss: 0.000000 Train Epoch: 2 [26240/104000 (25%)] Loss: 0.000000 Train Epoch: 2 [26880/104000 (26%)] Loss: 0.000000 Train Epoch: 2 [27520/104000 (26%)] Loss: 0.000000 Train Epoch: 2 [28160/104000 (27%)] Loss: 0.000000 Train Epoch: 2 [28800/104000 (28%)] Loss: 0.000000 Train Epoch: 2 [29440/104000 (28%)] Loss: 0.000000 Train Epoch: 2 [30080/104000 (29%)] Loss: 0.000000 Train Epoch: 2 [30720/104000 (30%)] Loss: 0.000000 Train Epoch: 2 [31360/104000 (30%)] Loss: 0.000000 Train Epoch: 2 [32000/104000 (31%)] Loss: 0.000000 Train Epoch: 2 [32640/104000 (31%)] Loss: 0.000000 Train Epoch: 2 [33280/104000 (32%)] Loss: 0.000000 Train Epoch: 2 [33920/104000 (33%)] Loss: 0.000000 Train Epoch: 2 [34560/104000 (33%)] Loss: 0.000000 Train Epoch: 2 [35200/104000 (34%)] Loss: 0.000000 Train Epoch: 2 [35840/104000 (34%)] Loss: 0.000000 Train Epoch: 2 [36480/104000 (35%)] Loss: 0.000000 Train Epoch: 2 [37120/104000 (36%)] Loss: 0.000000 Train Epoch: 2 [37760/104000 (36%)] Loss: 0.000000 Train Epoch: 2 [38400/104000 (37%)] Loss: 0.000001 Train Epoch: 2 [39040/104000 (38%)] Loss: 0.000000 Train Epoch: 2 [39680/104000 (38%)] Loss: 0.000000 Train Epoch: 2 [40320/104000 (39%)] Loss: 0.000000 Train Epoch: 2 [40960/104000 (39%)] Loss: 0.000000 Train Epoch: 2 [41600/104000 (40%)] Loss: 0.000000 Train Epoch: 2 [42240/104000 (41%)] Loss: 0.000000 Train Epoch: 2 [42880/104000 (41%)] Loss: 0.000000 Train Epoch: 2 [43520/104000 (42%)] Loss: 0.000000 Train Epoch: 2 [44160/104000 (42%)] Loss: 0.000000 Train Epoch: 2 [44800/104000 (43%)] Loss: 0.000000 Train Epoch: 2 [45440/104000 (44%)] Loss: 0.000000 Train Epoch: 2 [46080/104000 (44%)] Loss: 0.000000 Train Epoch: 2 [46720/104000 (45%)] Loss: 0.000000 Train Epoch: 2 [47360/104000 (46%)] Loss: 0.000000 Train Epoch: 2 [48000/104000 (46%)] Loss: 0.000000 Train Epoch: 2 [48640/104000 (47%)] Loss: 0.000000 Train Epoch: 2 [49280/104000 (47%)] Loss: 0.000000 Train Epoch: 2 [49920/104000 (48%)] Loss: 0.000000 Train Epoch: 2 [50560/104000 (49%)] Loss: 0.000000 Train Epoch: 2 [51200/104000 (49%)] Loss: 0.000000 Train Epoch: 2 [51840/104000 (50%)] Loss: 0.000000 Train Epoch: 2 [52480/104000 (50%)] Loss: 0.000000 Train Epoch: 2 [53120/104000 (51%)] Loss: 0.000000 Train Epoch: 2 [53760/104000 (52%)] Loss: 0.000000 Train Epoch: 2 [54400/104000 (52%)] Loss: 0.000000 Train Epoch: 2 [55040/104000 (53%)] Loss: 0.000000 Train Epoch: 2 [55680/104000 (54%)] Loss: 0.000000 Train Epoch: 2 [56320/104000 (54%)] Loss: 0.000000 Train Epoch: 2 [56960/104000 (55%)] Loss: 0.000000 Train Epoch: 2 [57600/104000 (55%)] Loss: 0.000000 Train Epoch: 2 [58240/104000 (56%)] Loss: 0.000000 Train Epoch: 2 [58880/104000 (57%)] Loss: 0.000000 Train Epoch: 2 [59520/104000 (57%)] Loss: 0.000000 Train Epoch: 2 [60160/104000 (58%)] Loss: 0.000000 Train Epoch: 2 [60800/104000 (58%)] Loss: 0.000000 Train Epoch: 2 [61440/104000 (59%)] Loss: 0.000000 Train Epoch: 2 [62080/104000 (60%)] Loss: 0.000000 Train Epoch: 2 [62720/104000 (60%)] Loss: 0.000000 Train Epoch: 2 [63360/104000 (61%)] Loss: 0.000000 Train Epoch: 2 [64000/104000 (62%)] Loss: 0.000000 Train Epoch: 2 [64640/104000 (62%)] Loss: 0.000000 Train Epoch: 2 [65280/104000 (63%)] Loss: 0.000000 Train Epoch: 2 [65920/104000 (63%)] Loss: 0.000000 Train Epoch: 2 [66560/104000 (64%)] Loss: 0.000000 Train Epoch: 2 [67200/104000 (65%)] Loss: 0.000000 Train Epoch: 2 [67840/104000 (65%)] Loss: 0.000000 Train Epoch: 2 [68480/104000 (66%)] Loss: 0.000000 Train Epoch: 2 [69120/104000 (66%)] Loss: 0.000000 Train Epoch: 2 [69760/104000 (67%)] Loss: 0.000000 Train Epoch: 2 [70400/104000 (68%)] Loss: 0.000000 Train Epoch: 2 [71040/104000 (68%)] Loss: 0.000000 Train Epoch: 2 [71680/104000 (69%)] Loss: 0.000000 Train Epoch: 2 [72320/104000 (70%)] Loss: 0.000000 Train Epoch: 2 [72960/104000 (70%)] Loss: 0.000000 Train Epoch: 2 [73600/104000 (71%)] Loss: 0.000000 Train Epoch: 2 [74240/104000 (71%)] Loss: 0.000000 Train Epoch: 2 [74880/104000 (72%)] Loss: 0.000000 Train Epoch: 2 [75520/104000 (73%)] Loss: 0.000000 Train Epoch: 2 [76160/104000 (73%)] Loss: 0.000000 Train Epoch: 2 [76800/104000 (74%)] Loss: 0.000000 Train Epoch: 2 [77440/104000 (74%)] Loss: 0.000000 Train Epoch: 2 [78080/104000 (75%)] Loss: 0.000000 Train Epoch: 2 [78720/104000 (76%)] Loss: 0.000000 Train Epoch: 2 [79360/104000 (76%)] Loss: 0.000000 Train Epoch: 2 [80000/104000 (77%)] Loss: 0.000000 Train Epoch: 2 [80640/104000 (78%)] Loss: 0.000000 Train Epoch: 2 [81280/104000 (78%)] Loss: 0.000000 Train Epoch: 2 [81920/104000 (79%)] Loss: 0.000000 Train Epoch: 2 [82560/104000 (79%)] Loss: 0.000000 Train Epoch: 2 [83200/104000 (80%)] Loss: 0.000000 Train Epoch: 2 [83840/104000 (81%)] Loss: 0.000000 Train Epoch: 2 [84480/104000 (81%)] Loss: 0.000000 Train Epoch: 2 [85120/104000 (82%)] Loss: 0.000000 Train Epoch: 2 [85760/104000 (82%)] Loss: 0.000000 Train Epoch: 2 [86400/104000 (83%)] Loss: 0.000000 Train Epoch: 2 [87040/104000 (84%)] Loss: 0.000000 Train Epoch: 2 [87680/104000 (84%)] Loss: 0.000000 Train Epoch: 2 [88320/104000 (85%)] Loss: 0.000000 Train Epoch: 2 [88960/104000 (86%)] Loss: 0.000000 Train Epoch: 2 [89600/104000 (86%)] Loss: 0.000000 Train Epoch: 2 [90240/104000 (87%)] Loss: 0.000001 Train Epoch: 2 [90880/104000 (87%)] Loss: 0.000000 Train Epoch: 2 [91520/104000 (88%)] Loss: 0.000000 Train Epoch: 2 [92160/104000 (89%)] Loss: 0.000000 Train Epoch: 2 [92800/104000 (89%)] Loss: 0.000000 Train Epoch: 2 [93440/104000 (90%)] Loss: 0.000000 Train Epoch: 2 [94080/104000 (90%)] Loss: 0.000000 Train Epoch: 2 [94720/104000 (91%)] Loss: 0.000000 Train Epoch: 2 [95360/104000 (92%)] Loss: 0.000000 Train Epoch: 2 [96000/104000 (92%)] Loss: 0.000000 Train Epoch: 2 [96640/104000 (93%)] Loss: 0.000000 Train Epoch: 2 [97280/104000 (94%)] Loss: 0.000000 Train Epoch: 2 [97920/104000 (94%)] Loss: 0.000000 Train Epoch: 2 [98560/104000 (95%)] Loss: 0.000000 Train Epoch: 2 [99200/104000 (95%)] Loss: 0.000000 Train Epoch: 2 [99840/104000 (96%)] Loss: 0.000000 Train Epoch: 2 [100480/104000 (97%)] Loss: 0.000000 Train Epoch: 2 [101120/104000 (97%)] Loss: 0.000000 Train Epoch: 2 [101760/104000 (98%)] Loss: 0.000000 Train Epoch: 2 [102400/104000 (98%)] Loss: 0.000000 Train Epoch: 2 [103040/104000 (99%)] Loss: 0.000000 Train Epoch: 2 [103680/104000 (100%)] Loss: 0.000000 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.7656, Accuracy: 25707/30000 (86%) Train Epoch: 3 [0/104000 (0%)] Loss: 0.000000 Train Epoch: 3 [640/104000 (1%)] Loss: 0.000000 Train Epoch: 3 [1280/104000 (1%)] Loss: 0.000000 Train Epoch: 3 [1920/104000 (2%)] Loss: 0.000000 Train Epoch: 3 [2560/104000 (2%)] Loss: 0.000000 Train Epoch: 3 [3200/104000 (3%)] Loss: 0.000000 Train Epoch: 3 [3840/104000 (4%)] Loss: 0.000010 Train Epoch: 3 [4480/104000 (4%)] Loss: 0.000000 Train Epoch: 3 [5120/104000 (5%)] Loss: 0.000000 Train Epoch: 3 [5760/104000 (6%)] Loss: 0.000000 Train Epoch: 3 [6400/104000 (6%)] Loss: 0.000000 Train Epoch: 3 [7040/104000 (7%)] Loss: 0.000000 Train Epoch: 3 [7680/104000 (7%)] Loss: 0.000000 Train Epoch: 3 [8320/104000 (8%)] Loss: 0.000000 Train Epoch: 3 [8960/104000 (9%)] Loss: 0.000000 Train Epoch: 3 [9600/104000 (9%)] Loss: 0.000000 Train Epoch: 3 [10240/104000 (10%)] Loss: 0.000000 Train Epoch: 3 [10880/104000 (10%)] Loss: 0.000000 Train Epoch: 3 [11520/104000 (11%)] Loss: 0.000000 Train Epoch: 3 [12160/104000 (12%)] Loss: 0.000000 Train Epoch: 3 [12800/104000 (12%)] Loss: 0.000000 Train Epoch: 3 [13440/104000 (13%)] Loss: 0.000000 Train Epoch: 3 [14080/104000 (14%)] Loss: 0.000000 Train Epoch: 3 [14720/104000 (14%)] Loss: 0.000000 Train Epoch: 3 [15360/104000 (15%)] Loss: 0.000000 Train Epoch: 3 [16000/104000 (15%)] Loss: 0.000000 Train Epoch: 3 [16640/104000 (16%)] Loss: 0.000000 Train Epoch: 3 [17280/104000 (17%)] Loss: 0.000000 Train Epoch: 3 [17920/104000 (17%)] Loss: 0.000000 Train Epoch: 3 [18560/104000 (18%)] Loss: 0.000000 Train Epoch: 3 [19200/104000 (18%)] Loss: 0.000000 Train Epoch: 3 [19840/104000 (19%)] Loss: 0.000000 Train Epoch: 3 [20480/104000 (20%)] Loss: 0.000000 Train Epoch: 3 [21120/104000 (20%)] Loss: 0.000000 Train Epoch: 3 [21760/104000 (21%)] Loss: 0.000000 Train Epoch: 3 [22400/104000 (22%)] Loss: 0.000000 Train Epoch: 3 [23040/104000 (22%)] Loss: 0.000000 Train Epoch: 3 [23680/104000 (23%)] Loss: 0.000000 Train Epoch: 3 [24320/104000 (23%)] Loss: 0.000000 Train Epoch: 3 [24960/104000 (24%)] Loss: 0.000000 Train Epoch: 3 [25600/104000 (25%)] Loss: 0.000000 Train Epoch: 3 [26240/104000 (25%)] Loss: 0.000000 Train Epoch: 3 [26880/104000 (26%)] Loss: 0.000000 Train Epoch: 3 [27520/104000 (26%)] Loss: 0.000000 Train Epoch: 3 [28160/104000 (27%)] Loss: 0.000000 Train Epoch: 3 [28800/104000 (28%)] Loss: 0.000000 Train Epoch: 3 [29440/104000 (28%)] Loss: 0.000000 Train Epoch: 3 [30080/104000 (29%)] Loss: 0.000000 Train Epoch: 3 [30720/104000 (30%)] Loss: 0.000000 Train Epoch: 3 [31360/104000 (30%)] Loss: 0.000000 Train Epoch: 3 [32000/104000 (31%)] Loss: 0.000000 Train Epoch: 3 [32640/104000 (31%)] Loss: 0.000000 Train Epoch: 3 [33280/104000 (32%)] Loss: 0.000000 Train Epoch: 3 [33920/104000 (33%)] Loss: 0.000002 Train Epoch: 3 [34560/104000 (33%)] Loss: 0.000000 Train Epoch: 3 [35200/104000 (34%)] Loss: 0.000000 Train Epoch: 3 [35840/104000 (34%)] Loss: 0.000000 Train Epoch: 3 [36480/104000 (35%)] Loss: 0.000000 Train Epoch: 3 [37120/104000 (36%)] Loss: 0.000000 Train Epoch: 3 [37760/104000 (36%)] Loss: 0.000000 Train Epoch: 3 [38400/104000 (37%)] Loss: 0.000002 Train Epoch: 3 [39040/104000 (38%)] Loss: 0.000000 Train Epoch: 3 [39680/104000 (38%)] Loss: 0.000000 Train Epoch: 3 [40320/104000 (39%)] Loss: 0.000000 Train Epoch: 3 [40960/104000 (39%)] Loss: 0.000000 Train Epoch: 3 [41600/104000 (40%)] Loss: 0.000000 Train Epoch: 3 [42240/104000 (41%)] Loss: 0.000000 Train Epoch: 3 [42880/104000 (41%)] Loss: 0.000000 Train Epoch: 3 [43520/104000 (42%)] Loss: 0.000000 Train Epoch: 3 [44160/104000 (42%)] Loss: 0.000000 Train Epoch: 3 [44800/104000 (43%)] Loss: 0.000000 Train Epoch: 3 [45440/104000 (44%)] Loss: 0.000000 Train Epoch: 3 [46080/104000 (44%)] Loss: 0.000000 Train Epoch: 3 [46720/104000 (45%)] Loss: 0.000000 Train Epoch: 3 [47360/104000 (46%)] Loss: 0.000000 Train Epoch: 3 [48000/104000 (46%)] Loss: 0.000000 Train Epoch: 3 [48640/104000 (47%)] Loss: 0.000000 Train Epoch: 3 [49280/104000 (47%)] Loss: 0.000000 Train Epoch: 3 [49920/104000 (48%)] Loss: 0.000000 Train Epoch: 3 [50560/104000 (49%)] Loss: 0.000000 Train Epoch: 3 [51200/104000 (49%)] Loss: 0.000000 Train Epoch: 3 [51840/104000 (50%)] Loss: 0.000000 Train Epoch: 3 [52480/104000 (50%)] Loss: 0.000000 Train Epoch: 3 [53120/104000 (51%)] Loss: 0.000000 Train Epoch: 3 [53760/104000 (52%)] Loss: 0.000000 Train Epoch: 3 [54400/104000 (52%)] Loss: 0.000000 Train Epoch: 3 [55040/104000 (53%)] Loss: 0.000000 Train Epoch: 3 [55680/104000 (54%)] Loss: 0.000000 Train Epoch: 3 [56320/104000 (54%)] Loss: 0.000035 Train Epoch: 3 [56960/104000 (55%)] Loss: 0.000000 Train Epoch: 3 [57600/104000 (55%)] Loss: 0.000000 Train Epoch: 3 [58240/104000 (56%)] Loss: 0.000000 Train Epoch: 3 [58880/104000 (57%)] Loss: 0.000000 Train Epoch: 3 [59520/104000 (57%)] Loss: 0.000000 Train Epoch: 3 [60160/104000 (58%)] Loss: 0.000000 Train Epoch: 3 [60800/104000 (58%)] Loss: 0.000000 Train Epoch: 3 [61440/104000 (59%)] Loss: 0.000000 Train Epoch: 3 [62080/104000 (60%)] Loss: 0.000000 Train Epoch: 3 [62720/104000 (60%)] Loss: 0.000000 Train Epoch: 3 [63360/104000 (61%)] Loss: 0.000552 Train Epoch: 3 [64000/104000 (62%)] Loss: 0.000000 Train Epoch: 3 [64640/104000 (62%)] Loss: 0.000000 Train Epoch: 3 [65280/104000 (63%)] Loss: 0.000000 Train Epoch: 3 [65920/104000 (63%)] Loss: 0.000000 Train Epoch: 3 [66560/104000 (64%)] Loss: 0.000000 Train Epoch: 3 [67200/104000 (65%)] Loss: 0.000000 Train Epoch: 3 [67840/104000 (65%)] Loss: 0.000000 Train Epoch: 3 [68480/104000 (66%)] Loss: 0.000000 Train Epoch: 3 [69120/104000 (66%)] Loss: 0.000000 Train Epoch: 3 [69760/104000 (67%)] Loss: 0.000000 Train Epoch: 3 [70400/104000 (68%)] Loss: 0.000000 Train Epoch: 3 [71040/104000 (68%)] Loss: 0.000000 Train Epoch: 3 [71680/104000 (69%)] Loss: 0.000000 Train Epoch: 3 [72320/104000 (70%)] Loss: 0.000000 Train Epoch: 3 [72960/104000 (70%)] Loss: 0.000000 Train Epoch: 3 [73600/104000 (71%)] Loss: 0.000000 Train Epoch: 3 [74240/104000 (71%)] Loss: 0.000000 Train Epoch: 3 [74880/104000 (72%)] Loss: 0.000000 Train Epoch: 3 [75520/104000 (73%)] Loss: 0.000000 Train Epoch: 3 [76160/104000 (73%)] Loss: 0.000000 Train Epoch: 3 [76800/104000 (74%)] Loss: 0.000000 Train Epoch: 3 [77440/104000 (74%)] Loss: 0.000000 Train Epoch: 3 [78080/104000 (75%)] Loss: 0.000000 Train Epoch: 3 [78720/104000 (76%)] Loss: 0.000000 Train Epoch: 3 [79360/104000 (76%)] Loss: 0.000000 Train Epoch: 3 [80000/104000 (77%)] Loss: 0.000000 Train Epoch: 3 [80640/104000 (78%)] Loss: 0.000000 Train Epoch: 3 [81280/104000 (78%)] Loss: 0.000000 Train Epoch: 3 [81920/104000 (79%)] Loss: 0.000000 Train Epoch: 3 [82560/104000 (79%)] Loss: 0.000000 Train Epoch: 3 [83200/104000 (80%)] Loss: 0.000000 Train Epoch: 3 [83840/104000 (81%)] Loss: 0.000000 Train Epoch: 3 [84480/104000 (81%)] Loss: 0.000000 Train Epoch: 3 [85120/104000 (82%)] Loss: 0.000000 Train Epoch: 3 [85760/104000 (82%)] Loss: 0.000000 Train Epoch: 3 [86400/104000 (83%)] Loss: 0.000000 Train Epoch: 3 [87040/104000 (84%)] Loss: 0.000000 Train Epoch: 3 [87680/104000 (84%)] Loss: 0.000000 Train Epoch: 3 [88320/104000 (85%)] Loss: 0.000000 Train Epoch: 3 [88960/104000 (86%)] Loss: 0.000000 Train Epoch: 3 [89600/104000 (86%)] Loss: 0.000003 Train Epoch: 3 [90240/104000 (87%)] Loss: 0.000000 Train Epoch: 3 [90880/104000 (87%)] Loss: 0.000000 Train Epoch: 3 [91520/104000 (88%)] Loss: 0.000000 Train Epoch: 3 [92160/104000 (89%)] Loss: 0.000000 Train Epoch: 3 [92800/104000 (89%)] Loss: 0.000000 Train Epoch: 3 [93440/104000 (90%)] Loss: 0.000000 Train Epoch: 3 [94080/104000 (90%)] Loss: 0.000000 Train Epoch: 3 [94720/104000 (91%)] Loss: 0.000000 Train Epoch: 3 [95360/104000 (92%)] Loss: 0.000000 Train Epoch: 3 [96000/104000 (92%)] Loss: 0.000000 Train Epoch: 3 [96640/104000 (93%)] Loss: 0.000000 Train Epoch: 3 [97280/104000 (94%)] Loss: 0.000000 Train Epoch: 3 [97920/104000 (94%)] Loss: 0.000000 Train Epoch: 3 [98560/104000 (95%)] Loss: 0.000000 Train Epoch: 3 [99200/104000 (95%)] Loss: 0.000000 Train Epoch: 3 [99840/104000 (96%)] Loss: 0.000000 Train Epoch: 3 [100480/104000 (97%)] Loss: 0.000000 Train Epoch: 3 [101120/104000 (97%)] Loss: 0.000000 Train Epoch: 3 [101760/104000 (98%)] Loss: 0.000000 Train Epoch: 3 [102400/104000 (98%)] Loss: 0.000000 Train Epoch: 3 [103040/104000 (99%)] Loss: 0.000000 Train Epoch: 3 [103680/104000 (100%)] Loss: 0.000000 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.6615, Accuracy: 26003/30000 (87%) Train Epoch: 4 [0/104000 (0%)] Loss: 0.000000 Train Epoch: 4 [640/104000 (1%)] Loss: 0.000000 Train Epoch: 4 [1280/104000 (1%)] Loss: 0.000000 Train Epoch: 4 [1920/104000 (2%)] Loss: 0.000000 Train Epoch: 4 [2560/104000 (2%)] Loss: 0.000000 Train Epoch: 4 [3200/104000 (3%)] Loss: 0.000000 Train Epoch: 4 [3840/104000 (4%)] Loss: 0.000000 Train Epoch: 4 [4480/104000 (4%)] Loss: 0.000000 Train Epoch: 4 [5120/104000 (5%)] Loss: 0.000000 Train Epoch: 4 [5760/104000 (6%)] Loss: 0.000000 Train Epoch: 4 [6400/104000 (6%)] Loss: 0.000000 Train Epoch: 4 [7040/104000 (7%)] Loss: 0.000000 Train Epoch: 4 [7680/104000 (7%)] Loss: 0.000000 Train Epoch: 4 [8320/104000 (8%)] Loss: 0.000000 Train Epoch: 4 [8960/104000 (9%)] Loss: 0.000000 Train Epoch: 4 [9600/104000 (9%)] Loss: 0.000000 Train Epoch: 4 [10240/104000 (10%)] Loss: 0.000000 Train Epoch: 4 [10880/104000 (10%)] Loss: 0.000000 Train Epoch: 4 [11520/104000 (11%)] Loss: 0.000000 Train Epoch: 4 [12160/104000 (12%)] Loss: 0.000000 Train Epoch: 4 [12800/104000 (12%)] Loss: 0.000000 Train Epoch: 4 [13440/104000 (13%)] Loss: 0.000000 Train Epoch: 4 [14080/104000 (14%)] Loss: 0.000000 Train Epoch: 4 [14720/104000 (14%)] Loss: 0.000000 Train Epoch: 4 [15360/104000 (15%)] Loss: 0.000000 Train Epoch: 4 [16000/104000 (15%)] Loss: 0.000000 Train Epoch: 4 [16640/104000 (16%)] Loss: 0.000000 Train Epoch: 4 [17280/104000 (17%)] Loss: 0.000000 Train Epoch: 4 [17920/104000 (17%)] Loss: 0.000000 Train Epoch: 4 [18560/104000 (18%)] Loss: 0.000000 Train Epoch: 4 [19200/104000 (18%)] Loss: 0.000007 Train Epoch: 4 [19840/104000 (19%)] Loss: 0.000000 Train Epoch: 4 [20480/104000 (20%)] Loss: 0.000000 Train Epoch: 4 [21120/104000 (20%)] Loss: 0.000000 Train Epoch: 4 [21760/104000 (21%)] Loss: 0.000000 Train Epoch: 4 [22400/104000 (22%)] Loss: 0.000000 Train Epoch: 4 [23040/104000 (22%)] Loss: 0.000000 Train Epoch: 4 [23680/104000 (23%)] Loss: 0.000000 Train Epoch: 4 [24320/104000 (23%)] Loss: 0.000000 Train Epoch: 4 [24960/104000 (24%)] Loss: 0.000000 Train Epoch: 4 [25600/104000 (25%)] Loss: 0.000000 Train Epoch: 4 [26240/104000 (25%)] Loss: 0.000000 Train Epoch: 4 [26880/104000 (26%)] Loss: 0.000000 Train Epoch: 4 [27520/104000 (26%)] Loss: 0.000000 Train Epoch: 4 [28160/104000 (27%)] Loss: 0.000000 Train Epoch: 4 [28800/104000 (28%)] Loss: 0.000000 Train Epoch: 4 [29440/104000 (28%)] Loss: 0.000000 Train Epoch: 4 [30080/104000 (29%)] Loss: 0.000000 Train Epoch: 4 [30720/104000 (30%)] Loss: 0.000000 Train Epoch: 4 [31360/104000 (30%)] Loss: 0.000000 Train Epoch: 4 [32000/104000 (31%)] Loss: 0.000000 Train Epoch: 4 [32640/104000 (31%)] Loss: 0.000000 Train Epoch: 4 [33280/104000 (32%)] Loss: 0.000000 Train Epoch: 4 [33920/104000 (33%)] Loss: 0.000000 Train Epoch: 4 [34560/104000 (33%)] Loss: 0.000000 Train Epoch: 4 [35200/104000 (34%)] Loss: 0.000000 Train Epoch: 4 [35840/104000 (34%)] Loss: 0.000000 Train Epoch: 4 [36480/104000 (35%)] Loss: 0.000000 Train Epoch: 4 [37120/104000 (36%)] Loss: 0.000000 Train Epoch: 4 [37760/104000 (36%)] Loss: 0.000000 Train Epoch: 4 [38400/104000 (37%)] Loss: 0.000000 Train Epoch: 4 [39040/104000 (38%)] Loss: 0.000000 Train Epoch: 4 [39680/104000 (38%)] Loss: 0.000000 Train Epoch: 4 [40320/104000 (39%)] Loss: 0.000000 Train Epoch: 4 [40960/104000 (39%)] Loss: 0.000000 Train Epoch: 4 [41600/104000 (40%)] Loss: 0.000000 Train Epoch: 4 [42240/104000 (41%)] Loss: 0.000000 Train Epoch: 4 [42880/104000 (41%)] Loss: 0.000000 Train Epoch: 4 [43520/104000 (42%)] Loss: 0.000000 Train Epoch: 4 [44160/104000 (42%)] Loss: 0.000000 Train Epoch: 4 [44800/104000 (43%)] Loss: 0.000000 Train Epoch: 4 [45440/104000 (44%)] Loss: 0.000000 Train Epoch: 4 [46080/104000 (44%)] Loss: 0.000000 Train Epoch: 4 [46720/104000 (45%)] Loss: 0.000000 Train Epoch: 4 [47360/104000 (46%)] Loss: 0.000000 Train Epoch: 4 [48000/104000 (46%)] Loss: 0.000000 Train Epoch: 4 [48640/104000 (47%)] Loss: 0.000000 Train Epoch: 4 [49280/104000 (47%)] Loss: 0.000000 Train Epoch: 4 [49920/104000 (48%)] Loss: 0.000000 Train Epoch: 4 [50560/104000 (49%)] Loss: 0.000000 Train Epoch: 4 [51200/104000 (49%)] Loss: 0.000000 Train Epoch: 4 [51840/104000 (50%)] Loss: 0.000000 Train Epoch: 4 [52480/104000 (50%)] Loss: 0.000000 Train Epoch: 4 [53120/104000 (51%)] Loss: 0.000000 Train Epoch: 4 [53760/104000 (52%)] Loss: 0.000000 Train Epoch: 4 [54400/104000 (52%)] Loss: 0.000001 Train Epoch: 4 [55040/104000 (53%)] Loss: 0.000000 Train Epoch: 4 [55680/104000 (54%)] Loss: 0.000000 Train Epoch: 4 [56320/104000 (54%)] Loss: 0.000000 Train Epoch: 4 [56960/104000 (55%)] Loss: 0.000000 Train Epoch: 4 [57600/104000 (55%)] Loss: 0.000003 Train Epoch: 4 [58240/104000 (56%)] Loss: 0.000000 Train Epoch: 4 [58880/104000 (57%)] Loss: 0.000000 Train Epoch: 4 [59520/104000 (57%)] Loss: 0.000000 Train Epoch: 4 [60160/104000 (58%)] Loss: 0.000000 Train Epoch: 4 [60800/104000 (58%)] Loss: 0.000000 Train Epoch: 4 [61440/104000 (59%)] Loss: 0.000000 Train Epoch: 4 [62080/104000 (60%)] Loss: 0.000000 Train Epoch: 4 [62720/104000 (60%)] Loss: 0.000000 Train Epoch: 4 [63360/104000 (61%)] Loss: 0.000000 Train Epoch: 4 [64000/104000 (62%)] Loss: 0.000000 Train Epoch: 4 [64640/104000 (62%)] Loss: 0.000000 Train Epoch: 4 [65280/104000 (63%)] Loss: 0.000000 Train Epoch: 4 [65920/104000 (63%)] Loss: 0.000000 Train Epoch: 4 [66560/104000 (64%)] Loss: 0.000000 Train Epoch: 4 [67200/104000 (65%)] Loss: 0.000000 Train Epoch: 4 [67840/104000 (65%)] Loss: 0.000000 Train Epoch: 4 [68480/104000 (66%)] Loss: 0.000000 Train Epoch: 4 [69120/104000 (66%)] Loss: 0.000000 Train Epoch: 4 [69760/104000 (67%)] Loss: 0.000000 Train Epoch: 4 [70400/104000 (68%)] Loss: 0.000000 Train Epoch: 4 [71040/104000 (68%)] Loss: 0.000000 Train Epoch: 4 [71680/104000 (69%)] Loss: 0.000000 Train Epoch: 4 [72320/104000 (70%)] Loss: 0.000000 Train Epoch: 4 [72960/104000 (70%)] Loss: 0.000000 Train Epoch: 4 [73600/104000 (71%)] Loss: 0.000000 Train Epoch: 4 [74240/104000 (71%)] Loss: 0.000000 Train Epoch: 4 [74880/104000 (72%)] Loss: 0.000000 Train Epoch: 4 [75520/104000 (73%)] Loss: 0.000000 Train Epoch: 4 [76160/104000 (73%)] Loss: 0.000000 Train Epoch: 4 [76800/104000 (74%)] Loss: 0.000000 Train Epoch: 4 [77440/104000 (74%)] Loss: 0.000000 Train Epoch: 4 [78080/104000 (75%)] Loss: 0.000000 Train Epoch: 4 [78720/104000 (76%)] Loss: 0.000000 Train Epoch: 4 [79360/104000 (76%)] Loss: 0.000000 Train Epoch: 4 [80000/104000 (77%)] Loss: 0.000000 Train Epoch: 4 [80640/104000 (78%)] Loss: 0.000000 Train Epoch: 4 [81280/104000 (78%)] Loss: 0.000000 Train Epoch: 4 [81920/104000 (79%)] Loss: 0.000000 Train Epoch: 4 [82560/104000 (79%)] Loss: 0.000000 Train Epoch: 4 [83200/104000 (80%)] Loss: 0.000000 Train Epoch: 4 [83840/104000 (81%)] Loss: 0.000000 Train Epoch: 4 [84480/104000 (81%)] Loss: 0.000000 Train Epoch: 4 [85120/104000 (82%)] Loss: 0.000000 Train Epoch: 4 [85760/104000 (82%)] Loss: 0.000000 Train Epoch: 4 [86400/104000 (83%)] Loss: 0.000000 Train Epoch: 4 [87040/104000 (84%)] Loss: 0.000000 Train Epoch: 4 [87680/104000 (84%)] Loss: 0.000000 Train Epoch: 4 [88320/104000 (85%)] Loss: 0.000000 Train Epoch: 4 [88960/104000 (86%)] Loss: 0.000000 Train Epoch: 4 [89600/104000 (86%)] Loss: 0.000000 Train Epoch: 4 [90240/104000 (87%)] Loss: 0.000000 Train Epoch: 4 [90880/104000 (87%)] Loss: 0.000000 Train Epoch: 4 [91520/104000 (88%)] Loss: 0.000000 Train Epoch: 4 [92160/104000 (89%)] Loss: 0.000000 Train Epoch: 4 [92800/104000 (89%)] Loss: 0.000000 Train Epoch: 4 [93440/104000 (90%)] Loss: 0.000000 Train Epoch: 4 [94080/104000 (90%)] Loss: 0.000000 Train Epoch: 4 [94720/104000 (91%)] Loss: 0.000000 Train Epoch: 4 [95360/104000 (92%)] Loss: 0.000000 Train Epoch: 4 [96000/104000 (92%)] Loss: 0.000000 Train Epoch: 4 [96640/104000 (93%)] Loss: 0.000000 Train Epoch: 4 [97280/104000 (94%)] Loss: 0.000000 Train Epoch: 4 [97920/104000 (94%)] Loss: 0.000000 Train Epoch: 4 [98560/104000 (95%)] Loss: 0.000000 Train Epoch: 4 [99200/104000 (95%)] Loss: 0.000000 Train Epoch: 4 [99840/104000 (96%)] Loss: 0.000000 Train Epoch: 4 [100480/104000 (97%)] Loss: 0.000000 Train Epoch: 4 [101120/104000 (97%)] Loss: 0.000000 Train Epoch: 4 [101760/104000 (98%)] Loss: 0.000000 Train Epoch: 4 [102400/104000 (98%)] Loss: 0.000000 Train Epoch: 4 [103040/104000 (99%)] Loss: 0.000000 Train Epoch: 4 [103680/104000 (100%)] Loss: 0.000000 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.4680, Accuracy: 26754/30000 (89%) Train Epoch: 5 [0/104000 (0%)] Loss: 0.000000 Train Epoch: 5 [640/104000 (1%)] Loss: 0.000000 Train Epoch: 5 [1280/104000 (1%)] Loss: 0.000000 Train Epoch: 5 [1920/104000 (2%)] Loss: 0.000000 Train Epoch: 5 [2560/104000 (2%)] Loss: 0.000000 Train Epoch: 5 [3200/104000 (3%)] Loss: 0.000000 Train Epoch: 5 [3840/104000 (4%)] Loss: 0.000000 Train Epoch: 5 [4480/104000 (4%)] Loss: 0.000000 Train Epoch: 5 [5120/104000 (5%)] Loss: 0.000000 Train Epoch: 5 [5760/104000 (6%)] Loss: 0.000000 Train Epoch: 5 [6400/104000 (6%)] Loss: 0.000000 Train Epoch: 5 [7040/104000 (7%)] Loss: 0.000000 Train Epoch: 5 [7680/104000 (7%)] Loss: 0.000033 Train Epoch: 5 [8320/104000 (8%)] Loss: 0.000000 Train Epoch: 5 [8960/104000 (9%)] Loss: 0.000000 Train Epoch: 5 [9600/104000 (9%)] Loss: 0.000000 Train Epoch: 5 [10240/104000 (10%)] Loss: 0.000000 Train Epoch: 5 [10880/104000 (10%)] Loss: 0.000000 Train Epoch: 5 [11520/104000 (11%)] Loss: 0.000000 Train Epoch: 5 [12160/104000 (12%)] Loss: 0.000000 Train Epoch: 5 [12800/104000 (12%)] Loss: 0.000000 Train Epoch: 5 [13440/104000 (13%)] Loss: 0.000000 Train Epoch: 5 [14080/104000 (14%)] Loss: 0.000000 Train Epoch: 5 [14720/104000 (14%)] Loss: 0.000000 Train Epoch: 5 [15360/104000 (15%)] Loss: 0.000000 Train Epoch: 5 [16000/104000 (15%)] Loss: 0.000000 Train Epoch: 5 [16640/104000 (16%)] Loss: 0.000000 Train Epoch: 5 [17280/104000 (17%)] Loss: 0.000000 Train Epoch: 5 [17920/104000 (17%)] Loss: 0.000000 Train Epoch: 5 [18560/104000 (18%)] Loss: 0.000000 Train Epoch: 5 [19200/104000 (18%)] Loss: 0.000000 Train Epoch: 5 [19840/104000 (19%)] Loss: 0.000000 Train Epoch: 5 [20480/104000 (20%)] Loss: 0.000000 Train Epoch: 5 [21120/104000 (20%)] Loss: 0.000000 Train Epoch: 5 [21760/104000 (21%)] Loss: 0.000000 Train Epoch: 5 [22400/104000 (22%)] Loss: 0.000000 Train Epoch: 5 [23040/104000 (22%)] Loss: 0.000000 Train Epoch: 5 [23680/104000 (23%)] Loss: 0.000000 Train Epoch: 5 [24320/104000 (23%)] Loss: 0.000000 Train Epoch: 5 [24960/104000 (24%)] Loss: 0.000000 Train Epoch: 5 [25600/104000 (25%)] Loss: 0.000000 Train Epoch: 5 [26240/104000 (25%)] Loss: 0.000000 Train Epoch: 5 [26880/104000 (26%)] Loss: 0.000000 Train Epoch: 5 [27520/104000 (26%)] Loss: 0.000000 Train Epoch: 5 [28160/104000 (27%)] Loss: 0.000000 Train Epoch: 5 [28800/104000 (28%)] Loss: 0.000000 Train Epoch: 5 [29440/104000 (28%)] Loss: 0.000000 Train Epoch: 5 [30080/104000 (29%)] Loss: 0.000000 Train Epoch: 5 [30720/104000 (30%)] Loss: 0.000000 Train Epoch: 5 [31360/104000 (30%)] Loss: 0.000000 Train Epoch: 5 [32000/104000 (31%)] Loss: 0.000000 Train Epoch: 5 [32640/104000 (31%)] Loss: 0.000000 Train Epoch: 5 [33280/104000 (32%)] Loss: 0.000000 Train Epoch: 5 [33920/104000 (33%)] Loss: 0.000001 Train Epoch: 5 [34560/104000 (33%)] Loss: 0.000000 Train Epoch: 5 [35200/104000 (34%)] Loss: 0.000000 Train Epoch: 5 [35840/104000 (34%)] Loss: 0.000000 Train Epoch: 5 [36480/104000 (35%)] Loss: 0.000000 Train Epoch: 5 [37120/104000 (36%)] Loss: 0.000000 Train Epoch: 5 [37760/104000 (36%)] Loss: 0.000000 Train Epoch: 5 [38400/104000 (37%)] Loss: 0.000000 Train Epoch: 5 [39040/104000 (38%)] Loss: 0.000000 Train Epoch: 5 [39680/104000 (38%)] Loss: 0.000000 Train Epoch: 5 [40320/104000 (39%)] Loss: 0.000000 Train Epoch: 5 [40960/104000 (39%)] Loss: 0.000000 Train Epoch: 5 [41600/104000 (40%)] Loss: 0.000000 Train Epoch: 5 [42240/104000 (41%)] Loss: 0.000000 Train Epoch: 5 [42880/104000 (41%)] Loss: 0.000000 Train Epoch: 5 [43520/104000 (42%)] Loss: 0.000000 Train Epoch: 5 [44160/104000 (42%)] Loss: 0.000000 Train Epoch: 5 [44800/104000 (43%)] Loss: 0.000000 Train Epoch: 5 [45440/104000 (44%)] Loss: 0.000000 Train Epoch: 5 [46080/104000 (44%)] Loss: 0.000000 Train Epoch: 5 [46720/104000 (45%)] Loss: 0.000000 Train Epoch: 5 [47360/104000 (46%)] Loss: 0.000000 Train Epoch: 5 [48000/104000 (46%)] Loss: 0.000000 Train Epoch: 5 [48640/104000 (47%)] Loss: 0.000000 Train Epoch: 5 [49280/104000 (47%)] Loss: 0.000000 Train Epoch: 5 [49920/104000 (48%)] Loss: 0.000000 Train Epoch: 5 [50560/104000 (49%)] Loss: 0.000000 Train Epoch: 5 [51200/104000 (49%)] Loss: 0.000000 Train Epoch: 5 [51840/104000 (50%)] Loss: 0.000000 Train Epoch: 5 [52480/104000 (50%)] Loss: 0.000000 Train Epoch: 5 [53120/104000 (51%)] Loss: 0.000000 Train Epoch: 5 [53760/104000 (52%)] Loss: 0.000000 Train Epoch: 5 [54400/104000 (52%)] Loss: 0.000000 Train Epoch: 5 [55040/104000 (53%)] Loss: 0.000000 Train Epoch: 5 [55680/104000 (54%)] Loss: 0.000000 Train Epoch: 5 [56320/104000 (54%)] Loss: 0.000000 Train Epoch: 5 [56960/104000 (55%)] Loss: 0.000000 Train Epoch: 5 [57600/104000 (55%)] Loss: 0.000000 Train Epoch: 5 [58240/104000 (56%)] Loss: 0.000000 Train Epoch: 5 [58880/104000 (57%)] Loss: 0.000000 Train Epoch: 5 [59520/104000 (57%)] Loss: 0.000000 Train Epoch: 5 [60160/104000 (58%)] Loss: 0.000000 Train Epoch: 5 [60800/104000 (58%)] Loss: 0.000000 Train Epoch: 5 [61440/104000 (59%)] Loss: 0.000000 Train Epoch: 5 [62080/104000 (60%)] Loss: 0.000000 Train Epoch: 5 [62720/104000 (60%)] Loss: 0.000000 Train Epoch: 5 [63360/104000 (61%)] Loss: 0.000000 Train Epoch: 5 [64000/104000 (62%)] Loss: 0.000000 Train Epoch: 5 [64640/104000 (62%)] Loss: 0.000000 Train Epoch: 5 [65280/104000 (63%)] Loss: 0.000000 Train Epoch: 5 [65920/104000 (63%)] Loss: 0.000000 Train Epoch: 5 [66560/104000 (64%)] Loss: 0.000000 Train Epoch: 5 [67200/104000 (65%)] Loss: 0.000000 Train Epoch: 5 [67840/104000 (65%)] Loss: 0.000000 Train Epoch: 5 [68480/104000 (66%)] Loss: 0.000000 Train Epoch: 5 [69120/104000 (66%)] Loss: 0.000000 Train Epoch: 5 [69760/104000 (67%)] Loss: 0.000000 Train Epoch: 5 [70400/104000 (68%)] Loss: 0.000000 Train Epoch: 5 [71040/104000 (68%)] Loss: 0.000000 Train Epoch: 5 [71680/104000 (69%)] Loss: 0.000000 Train Epoch: 5 [72320/104000 (70%)] Loss: 0.000000 Train Epoch: 5 [72960/104000 (70%)] Loss: 0.000000 Train Epoch: 5 [73600/104000 (71%)] Loss: 0.000000 Train Epoch: 5 [74240/104000 (71%)] Loss: 0.000000 Train Epoch: 5 [74880/104000 (72%)] Loss: 0.000000 Train Epoch: 5 [75520/104000 (73%)] Loss: 0.000000 Train Epoch: 5 [76160/104000 (73%)] Loss: 0.000000 Train Epoch: 5 [76800/104000 (74%)] Loss: 0.000000 Train Epoch: 5 [77440/104000 (74%)] Loss: 0.000000 Train Epoch: 5 [78080/104000 (75%)] Loss: 0.000000 Train Epoch: 5 [78720/104000 (76%)] Loss: 0.000000 Train Epoch: 5 [79360/104000 (76%)] Loss: 0.000000 Train Epoch: 5 [80000/104000 (77%)] Loss: 0.000000 Train Epoch: 5 [80640/104000 (78%)] Loss: 0.000000 Train Epoch: 5 [81280/104000 (78%)] Loss: 0.000000 Train Epoch: 5 [81920/104000 (79%)] Loss: 0.000000 Train Epoch: 5 [82560/104000 (79%)] Loss: 0.000000 Train Epoch: 5 [83200/104000 (80%)] Loss: 0.000000 Train Epoch: 5 [83840/104000 (81%)] Loss: 0.000000 Train Epoch: 5 [84480/104000 (81%)] Loss: 0.000000 Train Epoch: 5 [85120/104000 (82%)] Loss: 0.000000 Train Epoch: 5 [85760/104000 (82%)] Loss: 0.000000 Train Epoch: 5 [86400/104000 (83%)] Loss: 0.000000 Train Epoch: 5 [87040/104000 (84%)] Loss: 0.000000 Train Epoch: 5 [87680/104000 (84%)] Loss: 0.000000 Train Epoch: 5 [88320/104000 (85%)] Loss: 0.000000 Train Epoch: 5 [88960/104000 (86%)] Loss: 0.000000 Train Epoch: 5 [89600/104000 (86%)] Loss: 0.000000 Train Epoch: 5 [90240/104000 (87%)] Loss: 0.000000 Train Epoch: 5 [90880/104000 (87%)] Loss: 0.000000 Train Epoch: 5 [91520/104000 (88%)] Loss: 0.000000 Train Epoch: 5 [92160/104000 (89%)] Loss: 0.000000 Train Epoch: 5 [92800/104000 (89%)] Loss: 0.000000 Train Epoch: 5 [93440/104000 (90%)] Loss: 0.000000 Train Epoch: 5 [94080/104000 (90%)] Loss: 0.000000 Train Epoch: 5 [94720/104000 (91%)] Loss: 0.000000 Train Epoch: 5 [95360/104000 (92%)] Loss: 0.000000 Train Epoch: 5 [96000/104000 (92%)] Loss: 0.000000 Train Epoch: 5 [96640/104000 (93%)] Loss: 0.000000 Train Epoch: 5 [97280/104000 (94%)] Loss: 0.000000 Train Epoch: 5 [97920/104000 (94%)] Loss: 0.000000 Train Epoch: 5 [98560/104000 (95%)] Loss: 0.000000 Train Epoch: 5 [99200/104000 (95%)] Loss: 0.000000 Train Epoch: 5 [99840/104000 (96%)] Loss: 0.000000 Train Epoch: 5 [100480/104000 (97%)] Loss: 0.000000 Train Epoch: 5 [101120/104000 (97%)] Loss: 0.000000 Train Epoch: 5 [101760/104000 (98%)] Loss: 0.000000 Train Epoch: 5 [102400/104000 (98%)] Loss: 0.000000 Train Epoch: 5 [103040/104000 (99%)] Loss: 0.000000 Train Epoch: 5 [103680/104000 (100%)] Loss: 0.000000 Test set: Average loss: 0.0000, Accuracy: 26000/26000 (100%) Critical set: Average loss: 0.7019, Accuracy: 25847/30000 (86%) [100.0, 100.0, 100.0, 100.0, 100.0] [88.56, 91.44666666666667, 90.16666666666667, 85.94, 86.15666666666667]
Plot Accuracy¶
Let us now plot the accuracy of our calculation. Notice that even with a convolutional layer of depth 1 (one set of weights and a single bias!) we can get a 100% accuracy on the test set. We do less well on the critical data (somewhere between 80-90%) with lots of fluctuations from training run to training to run. Again, this shows you the incredible power and (some of the limitations) of all these ML methods. If the dataset we care about (critical region) is not exactly the dataset we train on, our accuracy can be significantly diminished.
In [7]:
from matplotlib import pyplot as plt
## Print the result for different N
%matplotlib inline
plt.plot(N_array, test_array, 'r-*', label="test")
plt.plot(N_array, critical_array, 'b-s', label="critical")
plt.ylim(60,110)
plt.xlabel('Depth of hidden layer', fontsize=24)
plt.xticks(N_array)
plt.ylabel('Accuracy', fontsize=24)
plt.legend(loc='best', fontsize=24)
plt.tick_params(axis='both', which='major', labelsize=24)
plt.tight_layout()
plt.show()
Exercises¶
- Do Step 6: modify the hyperparameters to optimize performance for the specific data set
- The strides used above do not account for the periodic boundary conditions. Define your own stride function in PyTorch to incorporate periodic boundary conditions