

In [1]:

    
import pandas as pd
from os import environ



In [2]:

    
import keras









    



Using TensorFlow backend.

原始資料來源的 SQL，這是抽樣過的資料，當中也有一筆資料是修改過的，因為當天 Server 似乎出了一些問題，導至流量大幅下降



In [326]:

    
sql="""SELECT
  date , sum(clicks) as cc 
FROM
  TABLE_DATE_RANGE(grouptargeting.search_console_, TIMESTAMP('2016-01-01'), TIMESTAMP('2017-10-31'))
where keyword like '%日本%' or keyword like '%東京%' or keyword like '%大阪%' or keyword like '%關西%' 
group by date
order by date"""


# sql="""
# SELECT
#   date , sum(clicks) as cc 
# FROM
# (
# select keyword ,(INTEGER(date/100)*100+1) as date,clicks from
#   TABLE_DATE_RANGE(grouptargeting.search_console_, TIMESTAMP('2016-01-01'), CURRENT_TIMESTAMP())
# )
  
#   where keyword like '%訂票%' 
# group by date
# order by date ,cc desc

# """



In [327]:

    
# load and plot dataset
from pandas import read_csv
from pandas import datetime
from matplotlib import pyplot
import matplotlib.dates as mdates

%matplotlib notebook

# %matplotlib inline  

# load dataset
def parser(x):
    return datetime.strptime(x, '%Y%m%d')

series = pd.read_gbq(sql,project_id=environ['PROJECT_ID'], verbose=False, private_key=environ['GOOGLE_KEY'])#,header=0, parse_dates=[0], index_col='date', squeeze=True, date_parser=parser)
series['date'] = pd.to_datetime(series['date'],format='%Y%m%d')
series.index = series['date']
del series['date']

# summarize first few rows
print(series.head())









    



WARNING:googleapiclient.discovery_cache:file_cache is unavailable when using oauth2client >= 4.0.0
Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/__init__.py", line 36, in autodetect
    from google.appengine.api import memcache
ImportError: No module named 'google.appengine'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/file_cache.py", line 33, in <module>
    from oauth2client.contrib.locked_file import LockedFile
ImportError: No module named 'oauth2client.contrib.locked_file'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/file_cache.py", line 37, in <module>
    from oauth2client.locked_file import LockedFile
ImportError: No module named 'oauth2client.locked_file'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/__init__.py", line 41, in autodetect
    from . import file_cache
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/file_cache.py", line 41, in <module>
    'file_cache is unavailable when using oauth2client >= 4.0.0')
ImportError: file_cache is unavailable when using oauth2client >= 4.0.0






    



               cc
date             
2016-01-01  29331
2016-01-02  33816
2016-01-03  37319
2016-01-04  36548
2016-01-05  37090



In [328]:

    
fig, ax = pyplot.subplots(1)
fig.autofmt_xdate()
pyplot.plot(series)
xfmt = mdates.DateFormatter('%Y-%m-%d')
ax.xaxis.set_major_formatter(xfmt)
pyplot.show()

為什麼 2017/1/3 流量跳起來

https://info.dingtaxi.com/zhtw/2017/01/2017chinese_new_year_tickets_info/



In [329]:

    
sql="""SELECT
  date,keyword , sum(clicks) as cc 
FROM
  TABLE_DATE_RANGE(grouptargeting.search_console_, TIMESTAMP('2016-01-01'), CURRENT_TIMESTAMP())
where keyword like '%訂票%' and date = 20170103
group by date,keyword
order by date ,cc desc"""
pd.read_gbq(sql,project_id=environ['PROJECT_ID'], verbose=True, private_key=environ['GOOGLE_KEY'])









    



WARNING:googleapiclient.discovery_cache:file_cache is unavailable when using oauth2client >= 4.0.0
Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/__init__.py", line 36, in autodetect
    from google.appengine.api import memcache
ImportError: No module named 'google.appengine'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/file_cache.py", line 33, in <module>
    from oauth2client.contrib.locked_file import LockedFile
ImportError: No module named 'oauth2client.contrib.locked_file'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/file_cache.py", line 37, in <module>
    from oauth2client.locked_file import LockedFile
ImportError: No module named 'oauth2client.locked_file'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/__init__.py", line 41, in autodetect
    from . import file_cache
  File "/usr/local/lib/python3.5/dist-packages/googleapiclient/discovery_cache/file_cache.py", line 41, in <module>
    'file_cache is unavailable when using oauth2client >= 4.0.0')
ImportError: file_cache is unavailable when using oauth2client >= 4.0.0






    



Requesting query... ok.
Job ID: job_tMlU5lbFtZvEUS3r_JiAKALqFKZV
Query running...
Query done.
Cache hit.

Retrieving results...
Got 455 rows.

Total time taken 2.34 s.
Finished at 2017-11-18 04:57:34.






    Out[329]:







  
    
      
      date
      keyword
      cc
    
  
  
    
      0
      20170103
      台鐵春節訂票時間2017
      291
    
    
      1
      20170103
      台鐵過年訂票2017
      52
    
    
      2
      20170103
      台鐵春節訂票
      51
    
    
      3
      20170103
      台鐵 春節 訂票 2017
      49
    
    
      4
      20170103
      春節火車訂票2017
      48
    
    
      5
      20170103
      吉卜力美術館訂票
      47
    
    
      6
      20170103
      濟州航空訂票教學
      28
    
    
      7
      20170103
      濟州航空 訂票
      26
    
    
      8
      20170103
      台鐵春節訂票2017
      25
    
    
      9
      20170103
      濟州航空 訂票教學
      24
    
    
      10
      20170103
      2017台鐵過年訂票
      23
    
    
      11
      20170103
      三鷹之森吉卜力美術館 網路訂票
      17
    
    
      12
      20170103
      濟州航空訂票
      17
    
    
      13
      20170103
      2017台鐵春節訂票時間
      15
    
    
      14
      20170103
      台鐵 春節訂票
      13
    
    
      15
      20170103
      春節台鐵訂票
      12
    
    
      16
      20170103
      2017春節台鐵訂票
      11
    
    
      17
      20170103
      tgv訂票
      11
    
    
      18
      20170103
      台鐵春節訂票時間
      11
    
    
      19
      20170103
      阿里山小火車訂票系統
      9
    
    
      20
      20170103
      台鐵語音訂票技巧
      9
    
    
      21
      20170103
      歐洲之星 訂票
      9
    
    
      22
      20170103
      虎航訂票教學
      8
    
    
      23
      20170103
      台鐵2017春節訂票
      8
    
    
      24
      20170103
      2017台鐵訂票
      7
    
    
      25
      20170103
      紐約百老匯訂票
      7
    
    
      26
      20170103
      台鐵春節火車訂票2017
      6
    
    
      27
      20170103
      春節訂票
      6
    
    
      28
      20170103
      2017台鐵春節訂票
      6
    
    
      29
      20170103
      虎航 訂票教學
      5
    
    
      ...
      ...
      ...
      ...
    
    
      425
      20170103
      全日空訂票
      0
    
    
      426
      20170103
      高鐵訂票早鳥優惠
      0
    
    
      427
      20170103
      高鐵早鳥訂票時間
      0
    
    
      428
      20170103
      台鐵電話訂票電話
      0
    
    
      429
      20170103
      奮起湖火車訂票
      0
    
    
      430
      20170103
      阿里山火車 訂票
      0
    
    
      431
      20170103
      奇美博物館訂票
      0
    
    
      432
      20170103
      歐洲鐵路訂票
      0
    
    
      433
      20170103
      火車訂票
      0
    
    
      434
      20170103
      高鐵網路取消訂票
      0
    
    
      435
      20170103
      維也納 金色大廳 訂票
      0
    
    
      436
      20170103
      台鐵訂票幾天前
      0
    
    
      437
      20170103
      高鐵 自由座 訂票
      0
    
    
      438
      20170103
      普悠瑪 訂票
      0
    
    
      439
      20170103
      森林小火車訂票
      0
    
    
      440
      20170103
      喜樂時代 訂票
      0
    
    
      441
      20170103
      高鐵 訂票 幾天前
      0
    
    
      442
      20170103
      花蓮火車訂票
      0
    
    
      443
      20170103
      國泰航空 訂票 教學
      0
    
    
      444
      20170103
      高鐵自由座訂票
      0
    
    
      445
      20170103
      聖家堂訂票
      0
    
    
      446
      20170103
      itx訂票
      0
    
    
      447
      20170103
      全日航 訂票
      0
    
    
      448
      20170103
      台東航空站訂票
      0
    
    
      449
      20170103
      高鐵網路訂票
      0
    
    
      450
      20170103
      韓巢訂票
      0
    
    
      451
      20170103
      jr北海道訂票
      0
    
    
      452
      20170103
      酷航app訂票
      0
    
    
      453
      20170103
      ekinet 訂票
      0
    
    
      454
      20170103
      香草航空 訂票
      0
    
  

455 rows × 3 columns

進行 scale to 0-1 ，方便作為 input 及 output (因為 sigmoid 介於 0~1 之間)



In [330]:

    
from sklearn.preprocessing import scale,MinMaxScaler
scaler = MinMaxScaler()



In [331]:

    
x = series.values



In [332]:

    
x = x.reshape([x.shape[0],1])



In [333]:

    
scaler.fit(x)









    



/usr/local/lib/python3.5/dist-packages/sklearn/utils/validation.py:475: DataConversionWarning: Data with input dtype int64 was converted to float64 by MinMaxScaler.
  warnings.warn(msg, DataConversionWarning)






    Out[333]:





MinMaxScaler(copy=True, feature_range=(0, 1))



In [334]:

    
x_scaled = scaler.transform(x)



In [335]:

    
fig, ax = pyplot.subplots(1)
fig.autofmt_xdate()
pyplot.plot(series.index,x_scaled)
xfmt = mdates.DateFormatter('%Y-%m-%d')
ax.xaxis.set_major_formatter(xfmt)
pyplot.show()

產生 x,y pair

舉列來說假設將 Step Size 設為 4 天，故一筆 Training Data ，為連續 4 天的流量。再來利用這４天的資料來預測第 5 天的流量
綠色的部是 Training Data(前4天的資料)，藍色的部份是需要被預測的部份。示意如下圖



In [336]:

    
#往回看 30 天前的每一筆資料
step_size = 15



In [337]:

    
date_list = series.index
date_list = date_list[step_size-1:-1]



In [338]:

    
print("原始資料長度:{}".format(x_scaled.shape))









    



原始資料長度:(659, 1)



In [339]:

    
def window_stack(a, stepsize=1, width=3):
    return np.hstack( a[i:1+i-width or None:stepsize] for i in range(0,width) )

import numpy as np
train_x = window_stack(x_scaled, stepsize=1, width=step_size)



In [340]:

    
# 最後一筆資料要放棄，因為沒有未來的答案作驗證

train_x = train_x[:-1]
train_x.shape









    Out[340]:





(644, 15)



In [341]:

    
# 請注意千萬不將每一筆(Row) 當中的最後一天資料作為 Training Data 中的 Input Data
train_y = np.array([i for i in x_scaled[step_size:]])

確認產出來的 Training Data 沒有包含到 Testing Data



In [342]:

    
train_y.shape









    Out[342]:





(644, 1)



In [343]:

    
train_x[0]









    Out[343]:





array([ 0.15089805,  0.17818638,  0.19949987,  0.19480883,  0.19810655,
        0.23921244,  0.18453236,  0.16873129,  0.18871839,  0.09144783,
        0.32480104,  0.31651415,  0.30774051,  0.30362749,  0.28347611])



In [344]:

    
train_x[1]









    Out[344]:





array([ 0.17818638,  0.19949987,  0.19480883,  0.19810655,  0.23921244,
        0.18453236,  0.16873129,  0.18871839,  0.09144783,  0.32480104,
        0.31651415,  0.30774051,  0.30362749,  0.28347611,  0.29874784])



In [345]:

    
train_y[0]









    Out[345]:





array([ 0.29874784])

Design Graph



In [346]:

    
# reshape input to be [samples, time steps, features]
trainX = np.reshape(train_x, (train_x.shape[0], step_size, 1))



In [347]:

    
from keras import Sequential
from keras.layers import LSTM,Dense
# create and fit the LSTM network
model = Sequential()
# input_shape(step_size,feature_dim)
model.add(LSTM(100, input_shape=(step_size,1), unroll=True))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adam',metrics=['accuracy'])
model.summary()









    



_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
lstm_6 (LSTM)                (None, 100)               40800     
_________________________________________________________________
dense_6 (Dense)              (None, 1)                 101       
=================================================================
Total params: 40,901
Trainable params: 40,901
Non-trainable params: 0
_________________________________________________________________

最後60天的筆資料不要看



In [348]:

    
validation_size = 60
val_loss = []
loss = []



In [ ]:

    
for _ in range(500):
    history = model.fit(trainX[:-1*validation_size],
              train_y[:-1*validation_size],
              epochs=1,shuffle=False, 
              validation_data=(trainX[-1*validation_size:],
              train_y[-1*validation_size:]))
    
    loss.append(history.history['loss'])
    val_loss.append(history.history['val_loss'])
    model.reset_states()









    



Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 2s 4ms/step - loss: 0.0293 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 618us/step - loss: 0.0073 - acc: 0.0017 - val_loss: 0.0088 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 601us/step - loss: 0.0065 - acc: 0.0017 - val_loss: 0.0048 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 588us/step - loss: 0.0056 - acc: 0.0017 - val_loss: 0.0050 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 597us/step - loss: 0.0056 - acc: 0.0017 - val_loss: 0.0055 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 578us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0056 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 566us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0056 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 574us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0055 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0055 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0054 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0054 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 568us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0053 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 576us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0052 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 583us/step - loss: 0.0057 - acc: 0.0017 - val_loss: 0.0051 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 570us/step - loss: 0.0056 - acc: 0.0017 - val_loss: 0.0050 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0056 - acc: 0.0017 - val_loss: 0.0050 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0056 - acc: 0.0017 - val_loss: 0.0049 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 594us/step - loss: 0.0056 - acc: 0.0017 - val_loss: 0.0049 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0056 - acc: 0.0017 - val_loss: 0.0048 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 573us/step - loss: 0.0056 - acc: 0.0017 - val_loss: 0.0047 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 580us/step - loss: 0.0055 - acc: 0.0017 - val_loss: 0.0047 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 568us/step - loss: 0.0055 - acc: 0.0017 - val_loss: 0.0046 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0055 - acc: 0.0017 - val_loss: 0.0046 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 574us/step - loss: 0.0055 - acc: 0.0017 - val_loss: 0.0046 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 572us/step - loss: 0.0055 - acc: 0.0017 - val_loss: 0.0045 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 571us/step - loss: 0.0055 - acc: 0.0017 - val_loss: 0.0045 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 569us/step - loss: 0.0055 - acc: 0.0017 - val_loss: 0.0044 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0044 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0044 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 568us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0043 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 585us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0043 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 568us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0043 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0042 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 568us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0042 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 586us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0042 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 572us/step - loss: 0.0054 - acc: 0.0017 - val_loss: 0.0041 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 570us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0041 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 571us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0041 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0041 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 586us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0040 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 573us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0040 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0040 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0040 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 581us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0039 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 572us/step - loss: 0.0053 - acc: 0.0017 - val_loss: 0.0039 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 571us/step - loss: 0.0052 - acc: 0.0017 - val_loss: 0.0039 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 556us/step - loss: 0.0052 - acc: 0.0017 - val_loss: 0.0039 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0052 - acc: 0.0017 - val_loss: 0.0039 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 566us/step - loss: 0.0052 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0052 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0052 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 569us/step - loss: 0.0052 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 572us/step - loss: 0.0051 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 570us/step - loss: 0.0051 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 560us/step - loss: 0.0051 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 570us/step - loss: 0.0051 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0051 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0051 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 588us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 572us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 582us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 565us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 566us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 573us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 577us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 573us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 551us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 544us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 545us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0050 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 560us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 566us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 571us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 555us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 566us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 565us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 569us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 555us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 551us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 571us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 560us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 576us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 580us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 560us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 569us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 558us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 574us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 574us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 558us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0049 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 570us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 553us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 542us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 550us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 542us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0042 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 530us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 549us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 542us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 544us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 530us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0036 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 555us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 550us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 552us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 542us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0035 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0048 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 556us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 551us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 532us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 579us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 613us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 616us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 545us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 602us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 532us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 584us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0037 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 555us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 604us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 532us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 553us/step - loss: 0.0047 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 573us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 572us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 528us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 607us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 545us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 566us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 605us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 603us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 573us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 556us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 568us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 531us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 542us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 526us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 530us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 550us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0038 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0046 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0045 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 551us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 545us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 542us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 530us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 530us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 549us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 544us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 556us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 570us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 581us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 550us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 565us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 545us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 530us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 531us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 544us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0032 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 549us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 589us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 557us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 531us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0033 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 529us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 532us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 560us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 545us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0034 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0044 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 545us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 550us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 531us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0031 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0022 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 529us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0042 - acc: 0.0017 - val_loss: 0.0030 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0043 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 550us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 555us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 553us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 565us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 560us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 557us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 569us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 602us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 576us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 566us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 568us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 581us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 578us/step - loss: 0.0041 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 573us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 583us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 561us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 557us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0023 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 555us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 558us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 557us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 556us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 557us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 567us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 551us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 565us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 584us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 557us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 556us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 563us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 573us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 557us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 553us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 560us/step - loss: 0.0040 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 571us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 565us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 551us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0024 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 528us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 532us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 562us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 564us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 569us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 531us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 539us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 553us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 528us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 530us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 572us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 560us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 544us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 544us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 570us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 544us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 615us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 582us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 575us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 628us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 604us/step - loss: 0.0039 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 530us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 630us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 571us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 610us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 531us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 614us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 580us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0025 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 528us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 537us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 549us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 550us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 554us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 546us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 535us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 541us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 551us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 555us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 540us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 568us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0038 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 533us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 544us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 543us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 549us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 549us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0026 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 524us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 538us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 532us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0027 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 547us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 536us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 532us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0029 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 559us/step - loss: 0.0037 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 534us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 548us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 550us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
584/584 [==============================] - 0s 590us/step - loss: 0.0036 - acc: 0.0017 - val_loss: 0.0028 - val_acc: 0.0000e+00
Train on 584 samples, validate on 60 samples
Epoch 1/1
 32/584 [>.............................] - ETA: 0s - loss: 0.0079 - acc: 0.0312

看一下 Error Rate 曲線



In [357]:

    
pyplot.figure()

pyplot.plot(loss)
pyplot.plot(val_loss)
pyplot.show()

看一下曲線擬合效果



In [358]:

    
predict_y = model.predict(trainX)



In [359]:

    
fig, ax = pyplot.subplots(1)
fig.autofmt_xdate()
xfmt = mdates.DateFormatter('%Y-%m-%d')
ax.xaxis.set_major_formatter(xfmt)


pyplot.plot(date_list, scaler.inverse_transform(predict_y))
pyplot.plot(date_list, scaler.inverse_transform(train_y))


pyplot.show()

來預測最後 60 天資料預出來的結果



In [360]:

    
predict_y = model.predict(trainX[-1*validation_size:])



In [361]:

    
predict_y = scaler.inverse_transform(predict_y)



In [362]:

    
predict_y.shape









    Out[362]:





(60, 1)



In [363]:

    
fig, ax = pyplot.subplots(1)
fig.autofmt_xdate()
xfmt = mdates.DateFormatter('%Y-%m-%d')
ax.xaxis.set_major_formatter(xfmt)

pyplot.plot(date_list[-60:],x[-1*(validation_size+1):-1])
pyplot.plot(date_list[-60:],predict_y)
pyplot.xticks(rotation=20)

pyplot.show()

來預測接下來 N 天的結果



In [364]:

    
hiistory_y = []
for i in range(60):
    s = (60-i)*-1
    e = (60-i)*-1 + 1
    if len(hiistory_y) > 0 :
        new_x = trainX[i].reshape(15).tolist()
        new_x[-1] = hiistory_y[-1]
        new_x = np.array(new_x)
        new_x = new_x.reshape([1,15,1])
        new_y = model.predict(new_x).reshape(1).tolist()[0]
    else:
        new_y = model.predict(trainX[s:e]).reshape(1).tolist()[0]
        
    hiistory_y.append(new_y)



In [365]:

    
fig, ax = pyplot.subplots(1)
fig.autofmt_xdate()
xfmt = mdates.DateFormatter('%Y-%m-%d')
ax.xaxis.set_major_formatter(xfmt)

pyplot.plot(date_list[-60:],x[-1*(validation_size+1):-1])


hiistory_y = scaler.inverse_transform([hiistory_y]).reshape([60])

pyplot.plot(date_list[-60:],hiistory_y)
pyplot.xticks(rotation=20)

pyplot.show()

心得觀察

LSTM 可以學習到 Period Pattern 是沒有問題的，但是似乎對於大幅的震盪以目前的 Model 來說無法完全的 Catch 到，但是還是有學到漲幅的趨勢預測
至於 LSTM 要如何調整震盪的幅度有兩個想法可以實驗看看
- 直接 Modified Training，將震盪幅度加大
- 修改 Loss Function ，把平方改為 2.x 次方不知道是否有效果



In [ ]:

	date	keyword	cc
0	20170103	台鐵春節訂票時間2017	291
1	20170103	台鐵過年訂票2017	52
2	20170103	台鐵春節訂票	51
3	20170103	台鐵春節訂票 2017	49
4	20170103	春節火車訂票2017	48
5	20170103	吉卜力美術館訂票	47
6	20170103	濟州航空訂票教學	28
7	20170103	濟州航空訂票	26
8	20170103	台鐵春節訂票2017	25
9	20170103	濟州航空訂票教學	24
10	20170103	2017台鐵過年訂票	23
11	20170103	三鷹之森吉卜力美術館網路訂票	17
12	20170103	濟州航空訂票	17
13	20170103	2017台鐵春節訂票時間	15
14	20170103	台鐵春節訂票	13
15	20170103	春節台鐵訂票	12
16	20170103	2017春節台鐵訂票	11
17	20170103	tgv訂票	11
18	20170103	台鐵春節訂票時間	11
19	20170103	阿里山小火車訂票系統	9
20	20170103	台鐵語音訂票技巧	9
21	20170103	歐洲之星訂票	9
22	20170103	虎航訂票教學	8
23	20170103	台鐵2017春節訂票	8
24	20170103	2017台鐵訂票	7
25	20170103	紐約百老匯訂票	7
26	20170103	台鐵春節火車訂票2017	6
27	20170103	春節訂票	6
28	20170103	2017台鐵春節訂票	6
29	20170103	虎航訂票教學	5
...	...	...	...
425	20170103	全日空訂票	0
426	20170103	高鐵訂票早鳥優惠	0
427	20170103	高鐵早鳥訂票時間	0
428	20170103	台鐵電話訂票電話	0
429	20170103	奮起湖火車訂票	0
430	20170103	阿里山火車訂票	0
431	20170103	奇美博物館訂票	0
432	20170103	歐洲鐵路訂票	0
433	20170103	火車訂票	0
434	20170103	高鐵網路取消訂票	0
435	20170103	維也納金色大廳訂票	0
436	20170103	台鐵訂票幾天前	0
437	20170103	高鐵自由座訂票	0
438	20170103	普悠瑪訂票	0
439	20170103	森林小火車訂票	0
440	20170103	喜樂時代訂票	0
441	20170103	高鐵訂票幾天前	0
442	20170103	花蓮火車訂票	0
443	20170103	國泰航空訂票教學	0
444	20170103	高鐵自由座訂票	0
445	20170103	聖家堂訂票	0
446	20170103	itx訂票	0
447	20170103	全日航訂票	0
448	20170103	台東航空站訂票	0
449	20170103	高鐵網路訂票	0
450	20170103	韓巢訂票	0
451	20170103	jr北海道訂票	0
452	20170103	酷航app訂票	0
453	20170103	ekinet 訂票	0
454	20170103	香草航空訂票	0