DietAI

Yapayzaka sinir ağları ile beslenme koçu oluşturma.

Okuldaki yapayzeka dersi için hazırladığım bu araştırmada temel anlamda bir beslenme koçu
oluşturmaya çalıştım. Projede temel hedefim günlük aldığım ve harcadığım kalorilerin yapayzeka ile
hesabını yapıp kilo alıp almayacağımı öğrenmek.

Malesef veri toplama ve yapay sinir ağı modelini oluşturma kısmında bazı sıkıntılar çektim.
Bu araştırmamı -yetersiz olsa da- bu alanda Türkçe çok az kaynak bulunduğu için paylaşmak istedim.
Umarım benden sonrakilere öncü olur.

Gerekli modüller

Modül adı	Kullanım alanı
tempfile	Geçici depolama
pandas	Veri işleme
tensorflow	Yapay sinir ağı
matplotlib	Veri görselleştirme

Kurulum

Öncelikle bu projeyi tensorflow yüklenmese dahi sanal ortamda çalıştırmanızı tavsiye ederim.
pip3 install -r requirements.txt komutu ile gerekli modülleri kurabilirsiniz. Eğer sorun çıkarsa
modüllerin her birinin sitesinde detaylı kurulumları mevcut.

Kütüphaneleri elle yüklemek isterseniz
  sudo pip3 install tensorflow
  sudo pip3 install pandas
  sudo apt-get install python3-matplotlib

Çalıştırma

jupyter-notebook DietAI.ipynb komutu ile bu dosyayı tarayıcı üstünde adım adım çalıştırabilirsiniz.
Daha çabuk sonuç almak istiyorsanız .py uzantılı dosyayı python3 ile çalıştırın.



In [1]:

    
import tempfile
import pandas as pd
import tensorflow as tf
import matplotlib.pyplot as plt
import random
import datetime
%matplotlib inline

Besin Verileri

Konumuz beslenme olduğu için haliyle besin verilerine ihtiyacımız var.
Aşağıda gördüğünüz ABBREV.csv dosyasında 92.000deb fazla besin kombinasyonunun detaylı bilgileri bulunmakta.



In [2]:

    
food_data = pd.DataFrame.from_csv("ABBREV.csv")
food_data.tail()









    Out[2]:






  
    
      
      Shrt_Desc
      Water_(g)
      Energ_Kcal
      Protein_(g)
      Lipid_Tot_(g)
      Ash_(g)
      Carbohydrt_(g)
      Fiber_TD_(g)
      Sugar_Tot_(g)
      Calcium_(mg)
      ...
      Vit_K_(µg)
      FA_Sat_(g)
      FA_Mono_(g)
      FA_Poly_(g)
      Cholestrl_(mg)
      GmWt_1
      GmWt_Desc1
      GmWt_2
      GmWt_Desc2
      Refuse_Pct
    
    
      No
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
    
  
  
    
      83110
      MACKEREL,SALTED
      43.00
      305
      18.50
      25.10
      13.40
      0.00
      0.0
      0.0
      66.0
      ...
      7.8
      7.148
      8.320
      6.210
      95.0
      136.0
      1 cup, cooked
      NaN
      NaN
      0.0
    
    
      90240
      SCALLOP,(BAY&SEA),CKD,STMD
      70.25
      111
      20.54
      0.84
      2.97
      5.41
      0.0
      0.0
      10.0
      ...
      0.0
      0.218
      0.082
      0.222
      41.0
      85.0
      3 oz
      NaN
      NaN
      0.0
    
    
      90480
      SYRUP,CANE
      26.00
      269
      0.00
      0.00
      0.86
      73.14
      0.0
      73.2
      13.0
      ...
      0.0
      0.000
      0.000
      0.000
      0.0
      NaN
      NaN
      NaN
      NaN
      0.0
    
    
      90560
      SNAIL,RAW
      79.20
      90
      16.10
      1.40
      1.30
      2.00
      0.0
      0.0
      10.0
      ...
      0.1
      0.361
      0.259
      0.252
      50.0
      85.0
      3 oz
      NaN
      NaN
      0.0
    
    
      93600
      TURTLE,GREEN,RAW
      78.50
      89
      19.80
      0.50
      1.20
      0.00
      0.0
      0.0
      118.0
      ...
      0.1
      0.127
      0.088
      0.170
      50.0
      85.0
      3 oz
      NaN
      NaN
      0.0
    
  

5 rows × 52 columns

Yapayzeka modelimizi eğitmek için kişilerin günlük aktiviteleri ile ilgili bilgilerine ihtiyacımız var.
Ben internette bulamadığım için kendi 'yapay' verilerimi oluşturdum.

Öncelikle beslenme alışkanlığını temsil etmek için 8 saat aralıkları ile 10.000 adet zaman verisini içeren
bir tablo oluşturdum. Bu tabloda yaş, kilo, boy, cinsiyet, günlük kalori ihtiyacı, günlük kalori alımı ve 
yakımı ile ilgili bilgiler tutulacak.

Modelle işimi kolaylaştırmak için tablodaki bazı kolonları sabit değerler olarak tanımladım.



In [3]:

    
rng = pd.date_range(datetime.datetime.now(), periods=10000, freq='24H')
personal_data = pd.DataFrame(index=rng)
personal_data['age'] = 23
personal_data['weight'] = 75
personal_data['height'] = 176
personal_data['gender'] = "male"



In [4]:

    
personal_data.head()









    Out[4]:






  
    
      
      age
      weight
      height
      gender
    
  
  
    
      2016-12-29 02:49:24.148388
      23
      75
      176
      male
    
    
      2016-12-30 02:49:24.148388
      23
      75
      176
      male
    
    
      2016-12-31 02:49:24.148388
      23
      75
      176
      male
    
    
      2017-01-01 02:49:24.148388
      23
      75
      176
      male
    
    
      2017-01-02 02:49:24.148388
      23
      75
      176
      male

Verilerin İşlenmesi

Besin tablosundan şimdilik işimize yarayacak olan 3 adet kolon var. 
Bunlar besin isimlerini içeren Shrt_Desc, kalori değerini içeren Energ_Kcal ve 
ölçüsünü belirten GmWt_1 kolonları.



In [5]:

    
food_data = food_data[['Shrt_Desc', 'Energ_Kcal', 'GmWt_1']]
food_data.head()









    Out[5]:






  
    
      
      Shrt_Desc
      Energ_Kcal
      GmWt_1
    
    
      No
      
      
      
    
  
  
    
      1001
      BUTTER,WITH SALT
      717
      227.00
    
    
      1002
      BUTTER,WHIPPED,WITH SALT
      717
      151.00
    
    
      1003
      BUTTER OIL,ANHYDROUS
      876
      205.00
    
    
      1004
      CHEESE,BLUE
      353
      28.35
    
    
      1005
      CHEESE,BRICK
      371
      132.00

Günlük kalori ihtiyacını hesaplamak için Harris–Benedict formülünü kullanacağız.
Bu formül için cinsiyet, kilo, boy ve yaş değerlerine ihtiyacımız var.
Formül kadınlar ve erkekler için ağaşıdaki gibi

$$BMRman = (10 × weight) + (6.25 × height) - (5 × age) + 5$$$$BMRwoman = (10 × weight) + (6.25 × height) - (5 × age) - 161$$

Egzersiz Yoğunluğu	Günlük Kilokalori İhtiyacı
Çok az veya hiç	BMR x 1.2
Hafif (haftada 1–3 gün)	BMR x 1.375
Orta (haftada 3–5 gün)	BMR x 1.55
Ağır(haftada 6–7 gün)	BMR x 1.725
Çok Ağır (günde 2 kez, ekstra antreman)	BMR x 1.9

Bu linkten daha fazla bilgi edinebilirsiniz.



In [6]:

    
activites = {"low":lambda x:x*1.2, 
             "light":lambda x:x*1.375, 
             "mid":lambda x:x*1.55, 
             "heavy":lambda x:x*1.725, 
             "very_heavy":lambda x:x*1.9}

def kcal(gender, weight, height, age, activity):
    """ Formüle göre BMR hesaplar ve aktiviteye göre günlük kcal değerini çıkartır"""
    if gender == "male":
        return  activites[activity]((10 * weight) + (6.25 * height) - (5 * age) + 5)
    else:
        return  activites[activity]((10 * weight) + (6.25 * height) - (5 * age) - 161)

Önce tabloya rastgele aktiviteler ekledik. Tablodaki her satıra erişip kcal fonksiyonuna gerekli parametreleri verdik.
Dönen değerlerden bir liste oluşturup bu listeyi kcal -günlük kkalori ihtiyacı- olarak tabloya ekledik



In [7]:

    
personal_data['activity'] = [random.choice(list(activites.keys())) for x in range(len(personal_data))]
personal_data['kcal']=[kcal(person[1]['gender'], person[1]['weight'], person[1]['height'], person[1]['age'], person[1]['activity'])
            for person in personal_data.iterrows()]
personal_data.head()









    Out[7]:






  
    
      
      age
      weight
      height
      gender
      activity
      kcal
    
  
  
    
      2016-12-29 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0
    
    
      2016-12-30 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0
    
    
      2016-12-31 02:49:24.148388
      23
      75
      176
      male
      heavy
      3001.5
    
    
      2017-01-01 02:49:24.148388
      23
      75
      176
      male
      low
      2088.0
    
    
      2017-01-02 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0

Eğitim verilerini oluşturmak için personal_data tablosundaki vatandaşa rastgele 3 öğün yemek yedireceğiz.
Yemek isimlerini ve kalori değerini de tabloya kolon olarak ekleyeceğiz. Bu işlem uzun sürebilir :)



In [8]:

    
personal_data['breakfast'] = [random.choice(list(food_data['Shrt_Desc']))  for x in range(len(personal_data))]
breakfast_kcal = [food_data.loc[food_data['Shrt_Desc'] == x, 'Energ_Kcal'] for x in list(personal_data['breakfast'])]
personal_data['breakfast_kcal'] = [k.iloc[0] for k in breakfast_kcal]

personal_data['lunch'] = [random.choice(list(food_data['Shrt_Desc']))  for x in range(len(personal_data))]
lunch_kcal = [food_data.loc[food_data['Shrt_Desc'] == x, 'Energ_Kcal'] for x in list(personal_data['lunch'])]
personal_data['lunch_kcal'] = [k.iloc[0] for k in lunch_kcal]

personal_data['dinner'] = [random.choice(list(food_data['Shrt_Desc']))  for x in range(len(personal_data))]
dinner_kcal = [food_data.loc[food_data['Shrt_Desc'] == x, 'Energ_Kcal'] for x in list(personal_data['dinner'])]
personal_data['dinner_kcal'] = [k.iloc[0] for k in dinner_kcal]

personal_data['total_kcal'] = personal_data['dinner_kcal'] + personal_data['lunch_kcal'] + personal_data['breakfast_kcal']

Ardından yediklerini yakması için her gün rastgele 1 saat egzersiz yaptıracağız. Elimizdeki egzersiz tablosunda egzersiz isimleri ve kilo başına dakikada yaktırdığı kkalori oranı bulunuyor.

Bu linkte detaylar mevcut.



In [9]:

    
exercise_data = pd.read_csv("activity_costs.csv", sep=";")
exercise_data.head()









    Out[9]:






  
    
      
      Activity
      Energy_Expenditure
    
  
  
    
      0
      Racquetball (recreational)
      0.07
    
    
      1
      Kayaking (leisure)
      0.04
    
    
      2
      Dancing (general)
      0.08
    
    
      3
      Golf (walking + bag)
      0.09
    
    
      4
      Running (5 mph, 12 min/mile)
      0.12



In [10]:

    
personal_data['workout'] = [random.choice(list(exercise_data['Activity']))  for x in range(len(personal_data))]
excercise_kcal = [exercise_data.loc[exercise_data['Activity'] == x, 'Energy_Expenditure'] for x in list(personal_data['workout'])]
personal_data['workout_kcal'] = [k.iloc[0] for k in excercise_kcal]
personal_data['workout_kcal'] = personal_data['workout_kcal']*personal_data['weight']*60 # energy_exp * weight * min
personal_data.head()









    Out[10]:






  
    
      
      age
      weight
      height
      gender
      activity
      kcal
      breakfast
      breakfast_kcal
      lunch
      lunch_kcal
      dinner
      dinner_kcal
      total_kcal
      workout
      workout_kcal
    
  
  
    
      2016-12-29 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0
      CANDIES,MARS SNACKFOOD US,SNICKERS CRUNCHER
      488
      RICE,WHITE,STMD,CHINESE RESTAURANT
      151
      SALAD DRSNG,PEPPERCORN DRSNG,COMM,REG
      564
      1203
      Dancing (general)
      360.0
    
    
      2016-12-30 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0
      ASPARAGUS,CND,REG PK,SOL&LIQUIDS
      15
      BEEF,GROUND,75% LN MEAT / 25% FAT,PATTY,CKD,PA...
      248
      BEEF,BRISKET,WHL,LN&FAT,1/8"FAT,ALL GRDS,CKD,BRSD
      331
      594
      Running (7.5 mph 8 min/mile)
      990.0
    
    
      2016-12-31 02:49:24.148388
      23
      75
      176
      male
      heavy
      3001.5
      FAST FD,PIZZA CHAIN,14" PIZZA,MEAT & VEG TOPPI...
      244
      FAST FOODS,MINIATURE CINN ROLLS
      403
      SWEET POTATO,CND,MSHD
      101
      748
      Sweeping
      225.0
    
    
      2017-01-01 02:49:24.148388
      23
      75
      176
      male
      low
      2088.0
      PORK,FRSH,VAR MEATS&BY-PRODUCTS,EARS,FRZ,RAW
      234
      CHEESE PRODUCT,PAST PROCESS,AMERICAN,RED FAT,F...
      240
      NUTS,PILINUTS,DRIED
      719
      1193
      Sitting Activities (very light)
      135.0
    
    
      2017-01-02 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0
      HORNED MELON (KIWANO)
      44
      MCDONALD'S,BIG 'N TASTY (WITHOUT MAYONNAISE)
      192
      MILK,DRY,NONFAT,INST,WO/ ADDED VIT A & VITAMIN D
      358
      594
      Cycling (stationary, 200W)
      810.0

Tabloya gerekli tüm verileri ekledik şimdi egzersizde harcanan enerjinin öğünlerden alınanlardan fazla olanları 1 olarak etiketleyeceğiz. Böylece yapay sinir ağı için eğitim datasetini hazırlamış olacağız.



In [11]:

    
def calibrate_labels(data_frame):
    """ calculates intake and cost value diff and adds it as label to dataframe """
    data_frame['label'] = data_frame[['total_kcal', 'workout_kcal']].apply(lambda x: x['workout_kcal'] - x['total_kcal'] > 0, axis=1).astype(int)
    return data_frame



In [12]:

    
personal_data = calibrate_labels(personal_data)
personal_data.to_csv("my_data.csv")
personal_data.head()









    Out[12]:






  
    
      
      age
      weight
      height
      gender
      activity
      kcal
      breakfast
      breakfast_kcal
      lunch
      lunch_kcal
      dinner
      dinner_kcal
      total_kcal
      workout
      workout_kcal
      label
    
  
  
    
      2016-12-29 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0
      CANDIES,MARS SNACKFOOD US,SNICKERS CRUNCHER
      488
      RICE,WHITE,STMD,CHINESE RESTAURANT
      151
      SALAD DRSNG,PEPPERCORN DRSNG,COMM,REG
      564
      1203
      Dancing (general)
      360.0
      0
    
    
      2016-12-30 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0
      ASPARAGUS,CND,REG PK,SOL&LIQUIDS
      15
      BEEF,GROUND,75% LN MEAT / 25% FAT,PATTY,CKD,PA...
      248
      BEEF,BRISKET,WHL,LN&FAT,1/8"FAT,ALL GRDS,CKD,BRSD
      331
      594
      Running (7.5 mph 8 min/mile)
      990.0
      1
    
    
      2016-12-31 02:49:24.148388
      23
      75
      176
      male
      heavy
      3001.5
      FAST FD,PIZZA CHAIN,14" PIZZA,MEAT & VEG TOPPI...
      244
      FAST FOODS,MINIATURE CINN ROLLS
      403
      SWEET POTATO,CND,MSHD
      101
      748
      Sweeping
      225.0
      0
    
    
      2017-01-01 02:49:24.148388
      23
      75
      176
      male
      low
      2088.0
      PORK,FRSH,VAR MEATS&BY-PRODUCTS,EARS,FRZ,RAW
      234
      CHEESE PRODUCT,PAST PROCESS,AMERICAN,RED FAT,F...
      240
      NUTS,PILINUTS,DRIED
      719
      1193
      Sitting Activities (very light)
      135.0
      0
    
    
      2017-01-02 02:49:24.148388
      23
      75
      176
      male
      very_heavy
      3306.0
      HORNED MELON (KIWANO)
      44
      MCDONALD'S,BIG 'N TASTY (WITHOUT MAYONNAISE)
      192
      MILK,DRY,NONFAT,INST,WO/ ADDED VIT A & VITAMIN D
      358
      594
      Cycling (stationary, 200W)
      810.0
      1



In [13]:

    
personal_data.columns









    Out[13]:





Index(['age', 'weight', 'height', 'gender', 'activity', 'kcal', 'breakfast',
       'breakfast_kcal', 'lunch', 'lunch_kcal', 'dinner', 'dinner_kcal',
       'total_kcal', 'workout', 'workout_kcal', 'label'],
      dtype='object')

Opsiyonel olarak tablodaki ilk n veriyi test verisi olarak kaydettik. 
İlerde bu programı tekrar çalıştırdığımızda veriler rastgele atanacağı için eski test verileri 
bizim için tutarlı ama farklı değerler olmuş olacak.

### !!! Bu hücre code türünde olmayabilir !!! personal_test = personal_data[:1000] # created already personal_test.to_csv("my_test.csv") personal_test.head()

Verileri Görselleştirelim

Günlük alınan ve harcanan kcal miktarı.



In [14]:

    
personal_data[['total_kcal', 'workout_kcal']][:20].plot(color=["r", "b"], figsize=(15, 4), fontsize=10, alpha=0.5)









    Out[14]:





<matplotlib.axes._subplots.AxesSubplot at 0x7fde892f7b00>

1. Pastada alınan ve verilen kilo oranlarını,
2. Pastada günlük kalori ihtiyaç türlerinin oranlarını çizdirdik.



In [15]:

    
personal_data['label'].value_counts().plot.pie(autopct="%.2f", colors=("cyan", "lightgreen"), figsize=(5, 5));
plt.figure();
personal_data['activity'].value_counts().plot.pie(autopct="%.2f", colors=("cyan", "lightgreen", "pink", "yellow", "lightblue"), figsize=(5, 5));

Öğünlerde alınan kkalorilerin kümülatif toplamı (tüm değerler sırayla toplanır).



In [16]:

    
personal_data[['breakfast_kcal', 'lunch_kcal', 'dinner_kcal', 'total_kcal']].cumsum().plot.area(figsize=(10,5), alpha=0.5)









    Out[16]:





<matplotlib.axes._subplots.AxesSubplot at 0x7fde8440deb8>

Yapay Sinir Ağı

Bu örnekte wide and deep learning tekniğini kullandık. 
Basitçe açıklamak gerekirse 'wide' olarak belirlediğimiz kısım ezberleme (Linear Classification),
'deep' olarak belirlediğimiz katmanlı sinir ağı ile sınıflandırma yapacağız.

Wide and deep learning tekniği büyük ölçekli regresyonlar (veriler arasındaki ilişkinin analizi) ve sparse (çok sayıda olası değerlerin bulunduğu veya kategorinin çok sayıda parçaya bölündüğü) sınıflandırma problemlerinde etkili.

Burada wide & deep learning ile ilgili güzel bir blog yazısı var. Burada da tensorflowun kendi örneği var.

Aşağıda sürekli ve kategorisel kolonları ve model için gerekli bazı değişkenleri belirledik.

Sürekli

* age
* weight
* height
* kcal
* breakfast_kcal
* lunch_kcal
* dinner_kcal
* total_kcal
* workout_kcal

Kategorsel

* gender
* activity
* breakfast
* lunch
* dinner
* workout



In [17]:

    
flags = tf.app.flags
FLAGS = flags.FLAGS

flags.DEFINE_string("model_dir", "", "Base directory for output models.")
flags.DEFINE_string("model_type", "wide_n_deep",
                    "Valid model types: {'wide', 'deep', 'wide_n_deep'}.")
flags.DEFINE_string(
    "train_data",
    "",
    "Path to the training data.")
flags.DEFINE_string(
    "test_data",
    "",
    "Path to the test data.")

COLUMNS = ['age', 'weight', 'height', 'gender', 'activity', 'kcal', 'breakfast',
       'breakfast_kcal', 'lunch', 'lunch_kcal', 'dinner', 'dinner_kcal',
       'total_kcal', 'workout', 'workout_kcal', 'label']
LABEL_COLUMN = "label"
CONTINUOUS_COLUMNS =  ['age', 'weight', 'height', 'kcal', 'breakfast_kcal', 'lunch_kcal', 'dinner_kcal', 'total_kcal', 'workout_kcal']
CATEGORICAL_COLUMNS= ["gender", 'activity', 'breakfast', 'lunch', 'dinner', 'workout']



In [18]:

    
def build_estimator(model_dir):
    """ Yapay sinir aği modeli."""
    # Sparse temel kolon
    gender = tf.contrib.layers.sparse_column_with_keys(column_name="gender",
                                                     keys=["female", "male"])
    
    # Surekli
    age = tf.contrib.layers.real_valued_column("age")
    weight = tf.contrib.layers.real_valued_column("weight")
    height = tf.contrib.layers.real_valued_column("height")
    kcal = tf.contrib.layers.real_valued_column("kcal")
    breakfast_kcal = tf.contrib.layers.real_valued_column("breakfast_kcal")
    lunch_kcal = tf.contrib.layers.real_valued_column("lunch_kcal")
    dinner_kcal = tf.contrib.layers.real_valued_column("dinner_kcal")
    total_kcal = tf.contrib.layers.real_valued_column("total_kcal")
    workout_kcal = tf.contrib.layers.real_valued_column("workout_kcal")

    # Kategorisel
    age_buckets = tf.contrib.layers.bucketized_column(age, boundaries=[
                                                        18, 25, 30, 35, 40, 45,
                                                        50, 55, 60, 65])
    
    activity = tf.contrib.layers.sparse_column_with_hash_bucket(
      "activity", hash_bucket_size=5)
    # öğünlerin türü belirlenebilirse hash_bucket_size daha tutarlı tanımlanabilir.
    breakfast = tf.contrib.layers.sparse_column_with_hash_bucket(
      "breakfast", hash_bucket_size=1000)
    lunch = tf.contrib.layers.sparse_column_with_hash_bucket(
      "lunch", hash_bucket_size=1000)
    dinner = tf.contrib.layers.sparse_column_with_hash_bucket(
      "dinner", hash_bucket_size=1000)
    
    workout = tf.contrib.layers.sparse_column_with_hash_bucket(
      "workout", hash_bucket_size=60) # 60 farklı egzersizimiz var
    
    # Wide columns and deep columns.
    wide_columns = [gender, age_buckets, activity, breakfast, lunch, dinner, workout]
    
    
    deep_columns = [
        tf.contrib.layers.embedding_column(gender, dimension=8),
        age,
        weight,
        height,
        kcal,
        breakfast_kcal,
        lunch_kcal,
        dinner_kcal,
        total_kcal,
        workout_kcal
    ]

    if FLAGS.model_type == "wide":
        m = tf.contrib.learn.LinearClassifier(model_dir=model_dir,
                                          feature_columns=wide_columns)
    elif FLAGS.model_type == "deep":
        m = tf.contrib.learn.DNNClassifier(
            model_dir=model_dir,
            feature_columns=deep_columns,
            hidden_units=[200, 100],  # gizli katma ve tensor sayısı
            optimizer=tflearn.optimizers.SGD(learning_rate=0.001), # Adam, SGD, RMSprop, RMSprop, Ftrl, AdaGrad 
            )
    else:
        m = tf.contrib.learn.DNNLinearCombinedClassifier(
            model_dir=model_dir,
            linear_feature_columns=wide_columns,
            dnn_feature_columns=deep_columns,
            dnn_hidden_units=[100, 50])
    return m



In [19]:

    
def input_fn(df):
    """ Kolonları tensorflow sabitlerine dönüştürür ve birleştirir. """
   
    continuous_cols = {k: tf.constant(df[k].values) for k in CONTINUOUS_COLUMNS}
    
    categorical_cols = {k: tf.SparseTensor(
        indices=[[i, 0] for i in range(df[k].size)],
        values=df[k].values,
        shape=[df[k].size, 1])
                      for k in CATEGORICAL_COLUMNS}
    
    feature_cols = dict(continuous_cols)
    feature_cols.update(categorical_cols)
    label = tf.constant(df[LABEL_COLUMN].values)

    return feature_cols, label

Geçici bir klasor içinde modeli oluşturduk.



In [20]:

    
model_dir = tempfile.mkdtemp()
m = build_estimator(model_dir)









    



WARNING:tensorflow:The default value of combiner will change from "sum" to "sqrtn" after 2016/11/01.
WARNING:tensorflow:The default value of combiner will change from "sum" to "sqrtn" after 2016/11/01.
WARNING:tensorflow:The default value of combiner will change from "sum" to "sqrtn" after 2016/11/01.
WARNING:tensorflow:The default value of combiner will change from "sum" to "sqrtn" after 2016/11/01.
WARNING:tensorflow:The default value of combiner will change from "sum" to "sqrtn" after 2016/11/01.
WARNING:tensorflow:The default value of combiner will change from "sum" to "sqrtn" after 2016/11/01.
WARNING:tensorflow:The default value of combiner will change from "mean" to "sqrtn" after 2016/11/01.
INFO:tensorflow:Using default config.
INFO:tensorflow:Using config: {'save_checkpoints_secs': 600, 'keep_checkpoint_every_n_hours': 10000, '_evaluation_master': '', '_task_type': None, '_num_ps_replicas': 0, 'save_checkpoints_steps': None, '_cluster_spec': <tensorflow.python.training.server_lib.ClusterSpec object at 0x7fde843bfc88>, 'keep_checkpoint_max': 5, 'save_summary_steps': 100, 'tf_random_seed': None, '_task_id': 0, '_environment': 'local', '_is_chief': True, '_master': '', 'tf_config': gpu_options {
  per_process_gpu_memory_fraction: 1
}
}

Hazırladığımız tabloyu input_fn fonksiyonu ile tensorflow için hazır hale getirdik ve modelimize
verdik. 2. parametre olarak modeli eğitme işleminin kaç kez gerçekleştirileceğini belirledik.



In [21]:

    
train_res = m.fit(input_fn=lambda: input_fn(personal_data), steps=2000)









    



WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:711 in fit.: calling BaseEstimator.fit (from tensorflow.contrib.learn.python.learn.estimators.estimator) with x is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:711 in fit.: calling BaseEstimator.fit (from tensorflow.contrib.learn.python.learn.estimators.estimator) with y is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:711 in fit.: calling BaseEstimator.fit (from tensorflow.contrib.learn.python.learn.estimators.estimator) with batch_size is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.
INFO:tensorflow:Create CheckpointSaverHook.
INFO:tensorflow:loss = 59.2865, step = 2
INFO:tensorflow:Saving checkpoints for 2 into /tmp/tmpxg2sbwit/model.ckpt.
WARNING:tensorflow:*******************************************************
WARNING:tensorflow:TensorFlow's V1 checkpoint format has been deprecated.
WARNING:tensorflow:Consider switching to the more efficient V2 format:
WARNING:tensorflow:   `tf.train.Saver(write_version=tf.train.SaverDef.V2)`
WARNING:tensorflow:now on by default.
WARNING:tensorflow:*******************************************************
INFO:tensorflow:global_step/sec: 4.82363
INFO:tensorflow:loss = 0.0411305, step = 202
INFO:tensorflow:global_step/sec: 6.30966
INFO:tensorflow:global_step/sec: 8.78707
INFO:tensorflow:loss = 0.0350631, step = 402
INFO:tensorflow:global_step/sec: 7.48301
INFO:tensorflow:global_step/sec: 9.36997
INFO:tensorflow:loss = 0.0254813, step = 602
INFO:tensorflow:global_step/sec: 10.0181
INFO:tensorflow:global_step/sec: 9.30019
INFO:tensorflow:loss = 0.0152094, step = 802
INFO:tensorflow:global_step/sec: 6.48101
INFO:tensorflow:global_step/sec: 8.75894
INFO:tensorflow:loss = 0.0111998, step = 1002
INFO:tensorflow:global_step/sec: 11.5281
INFO:tensorflow:global_step/sec: 11.3312
INFO:tensorflow:loss = 0.00973692, step = 1202
INFO:tensorflow:global_step/sec: 10.9574
INFO:tensorflow:global_step/sec: 11.6958
INFO:tensorflow:loss = 0.00879224, step = 1402
INFO:tensorflow:global_step/sec: 11.7021
INFO:tensorflow:global_step/sec: 11.7832
INFO:tensorflow:loss = 0.0081287, step = 1602
INFO:tensorflow:global_step/sec: 11.7502
INFO:tensorflow:global_step/sec: 11.5564
INFO:tensorflow:loss = 0.00765403, step = 1802
INFO:tensorflow:global_step/sec: 11.5464
INFO:tensorflow:global_step/sec: 11.5935
INFO:tensorflow:loss = 0.00729846, step = 2002
INFO:tensorflow:global_step/sec: 11.5952
INFO:tensorflow:Saving checkpoints for 2002 into /tmp/tmpxg2sbwit/model.ckpt.
WARNING:tensorflow:*******************************************************
WARNING:tensorflow:TensorFlow's V1 checkpoint format has been deprecated.
WARNING:tensorflow:Consider switching to the more efficient V2 format:
WARNING:tensorflow:   `tf.train.Saver(write_version=tf.train.SaverDef.V2)`
WARNING:tensorflow:now on by default.
WARNING:tensorflow:*******************************************************
INFO:tensorflow:Loss for final step: 0.00729846.

Önceden kaydettiğimiz test dosyasını (personal_data'ın bir parçası da olabilir) modelimize verdik
ve yine tekrar sayısını belirledik.



In [22]:

    
personal_test = pd.read_csv("my_test.csv")



In [23]:

    
results = m.evaluate(input_fn=lambda: input_fn(personal_test), steps=1000)









    



WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:719 in evaluate.: calling BaseEstimator.evaluate (from tensorflow.contrib.learn.python.learn.estimators.estimator) with x is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:719 in evaluate.: calling BaseEstimator.evaluate (from tensorflow.contrib.learn.python.learn.estimators.estimator) with y is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:719 in evaluate.: calling BaseEstimator.evaluate (from tensorflow.contrib.learn.python.learn.estimators.estimator) with batch_size is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.
INFO:tensorflow:Restored model from /tmp/tmpxg2sbwit
INFO:tensorflow:Eval steps [0,1000) for training step 2002.
INFO:tensorflow:Results after 10 steps (0.035 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 20 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 30 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 40 steps (0.047 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 50 steps (0.049 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 60 steps (0.027 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 70 steps (0.019 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 80 steps (0.017 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 90 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 100 steps (0.045 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 110 steps (0.043 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 120 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 130 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 140 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 150 steps (0.047 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 160 steps (0.021 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 170 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 180 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245705, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 190 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 200 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 210 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 220 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 230 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 240 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100907, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 250 steps (0.017 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 260 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 270 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 280 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 290 steps (0.017 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 300 steps (0.020 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 310 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 320 steps (0.028 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 330 steps (0.019 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 340 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 350 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 360 steps (0.027 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 370 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 380 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 390 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 400 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 410 steps (0.017 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 420 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 430 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 440 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 450 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 460 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 470 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 480 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 490 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 500 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 510 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 520 steps (0.018 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 530 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 540 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 550 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 560 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 570 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 580 steps (0.020 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 590 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 600 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 610 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 620 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 630 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 640 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 650 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 660 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 670 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 680 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 690 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 700 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 710 steps (0.019 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245704, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 720 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 730 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 740 steps (0.018 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 750 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 760 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 770 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 780 steps (0.019 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 790 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 800 steps (0.018 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 810 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 820 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 830 steps (0.030 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 840 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 850 steps (0.018 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 860 steps (0.051 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 870 steps (0.043 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 880 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 890 steps (0.025 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 900 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 910 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 920 steps (0.017 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 930 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 940 steps (0.016 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 950 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 960 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 970 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 980 steps (0.014 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 990 steps (0.015 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Results after 1000 steps (0.013 sec/batch): accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951.
INFO:tensorflow:Saving evaluation summary for step 2002: accuracy = 0.996, accuracy/baseline_label_mean = 0.247, accuracy/threshold_0.500000_mean = 0.996, auc = 0.999892, labels/actual_label_mean = 0.247, labels/prediction_mean = 0.245703, loss = 0.0100908, precision/positive_threshold_0.500000_mean = 0.987952, recall/positive_threshold_0.500000_mean = 0.995951

Son olarak test tablosundan 10 değer aldık ve tahmin etmesi için modele verdik.



In [24]:

    
sample = personal_test[10:20]
sample.head()









    Out[24]:






  
    
      
      Unnamed: 0
      age
      weight
      height
      gender
      activity
      kcal
      breakfast
      breakfast_kcal
      lunch
      lunch_kcal
      dinner
      dinner_kcal
      total_kcal
      workout
      workout_kcal
      label
    
  
  
    
      10
      2017-01-08 02:24:13.995011
      23
      75
      176
      male
      mid
      2697.0
      BEEF,CHUCK,CLOD,TOP BLADE,STEAK,LN & FAT,0" FA...
      176
      PUDDINGS,BANANA,DRY MIX,INST,W/ ADDED OIL
      386
      PEANUTS,ALL TYPES,RAW
      567
      1129
      Weeding/Gardening
      315.0
      0
    
    
      11
      2017-01-09 02:24:13.995011
      23
      75
      176
      male
      very_heavy
      3306.0
      CARROTS,CND,NO SALT,SOL&LIQUIDS
      23
      PACE,SALSA REFRIED BNS
      60
      FRANKFURTER,BF,UNHTD
      313
      396
      Running 5.5 mph (11 min/mile)
      630.0
      1
    
    
      12
      2017-01-10 02:24:13.995011
      23
      75
      176
      male
      heavy
      3001.5
      BABYFOOD,TEETHING BISCUITS
      391
      SESAME FLOUR,HIGH-FAT
      526
      KELLOGG'S EGGO LOWFAT BLUEBERRY NUTRI - GRAIN ...
      208
      1125
      Carrying Groceries (light)
      315.0
      0
    
    
      13
      2017-01-11 02:24:13.995011
      23
      75
      176
      male
      low
      2088.0
      PORK,FRSH,SPARERIBS,LN&FAT,CKD,BRSD
      397
      SAUCE,BARBECUE,SWT BABY RAY'S,ORIGINAL
      192
      KRAFT VELVEETA LT RED FAT PAST PROCESS CHS PRO...
      222
      811
      Chopping Wood
      405.0
      0
    
    
      14
      2017-01-12 02:24:13.995011
      23
      75
      176
      male
      very_heavy
      3306.0
      BEEF,TOP SIRLOIN,STEAK,LN & FAT,1/8" FAT,ALL G...
      201
      OIL,INDUSTRIAL,SOY,REFINED,FOR WOKS & LT FRYING
      884
      BABYFOOD,DINNER,VEG&CHICK,JR
      53
      1138
      Sexual Activity (kissing, hugs)
      90.0
      0



In [25]:

    
res = m.predict(input_fn=lambda:input_fn(sample))









    



WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:747 in predict.: calling BaseEstimator.predict (from tensorflow.contrib.learn.python.learn.estimators.estimator) with x is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:747 in predict.: calling BaseEstimator.predict (from tensorflow.contrib.learn.python.learn.estimators.estimator) with batch_size is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:747 in predict.: calling BaseEstimator.predict (from tensorflow.contrib.learn.python.learn.estimators.estimator) with as_iterable is deprecated and will be removed after 2016-12-01.
Instructions for updating:
Estimator is decoupled from Scikit Learn interface by moving into
separate class SKCompat. Arguments x, y and batch_size are only
available in the SKCompat class, Estimator will only accept input_fn.
Example conversion:
  est = Estimator(...) -> est = SKCompat(Estimator(...))
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:Rank of input Tensor (1) should be the same as output_rank (2) for column. Will attempt to expand dims. It is highly recommended that you resize your input, as this behavior may change.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:364 in _add_hidden_layer_summary.: scalar_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.scalar. Note that tf.summary.scalar uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on the scope they are created in. Also, passing a tensor or list of tags to a scalar summary op is no longer supported.
WARNING:tensorflow:From /usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py:365 in _add_hidden_layer_summary.: histogram_summary (from tensorflow.python.ops.logging_ops) is deprecated and will be removed after 2016-11-30.
Instructions for updating:
Please switch to tf.summary.histogram. Note that tf.summary.histogram uses the node name instead of the tag. This means that TensorFlow will automatically de-duplicate summary names based on their scope.

0.99599999% doğruluk oranı gayet başarılı :)



In [26]:

    
import pprint
pprint.pprint(results)
for gercek, tahmin in zip([x for x in sample['label']], [next(res) for x in range(10)]):
    print("Gercek: {}, Tahmin Edilen: {} ".format(gercek, tahmin))









    



{'accuracy': 0.99599999,
 'accuracy/baseline_label_mean': 0.24699999,
 'accuracy/threshold_0.500000_mean': 0.99599999,
 'auc': 0.99989247,
 'global_step': 2002,
 'labels/actual_label_mean': 0.24699999,
 'labels/prediction_mean': 0.24570338,
 'loss': 0.010090809,
 'precision/positive_threshold_0.500000_mean': 0.98795182,
 'recall/positive_threshold_0.500000_mean': 0.99595141}
INFO:tensorflow:Loading model from checkpoint: /tmp/tmpxg2sbwit/model.ckpt-2002-?????-of-00001.
Gercek: 0, Tahmin Edilen: 0 
Gercek: 1, Tahmin Edilen: 1 
Gercek: 0, Tahmin Edilen: 0 
Gercek: 0, Tahmin Edilen: 0 
Gercek: 0, Tahmin Edilen: 0 
Gercek: 1, Tahmin Edilen: 1 
Gercek: 0, Tahmin Edilen: 0 
Gercek: 1, Tahmin Edilen: 1 
Gercek: 1, Tahmin Edilen: 1 
Gercek: 0, Tahmin Edilen: 0



In [ ]:

	Shrt_Desc	Water_(g)	Energ_Kcal	Protein_(g)	Lipid_Tot_(g)	Ash_(g)	Carbohydrt_(g)	Fiber_TD_(g)	Sugar_Tot_(g)	Calcium_(mg)	...	Vit_K_(µg)	FA_Sat_(g)	FA_Mono_(g)	FA_Poly_(g)	Cholestrl_(mg)	GmWt_1	GmWt_Desc1	GmWt_2	GmWt_Desc2	Refuse_Pct
No
83110	MACKEREL,SALTED	43.00	305	18.50	25.10	13.40	0.00	0.0	0.0	66.0	...	7.8	7.148	8.320	6.210	95.0	136.0	1 cup, cooked	NaN	NaN	0.0
90240	SCALLOP,(BAY&SEA),CKD,STMD	70.25	111	20.54	0.84	2.97	5.41	0.0	0.0	10.0	...	0.0	0.218	0.082	0.222	41.0	85.0	3 oz	NaN	NaN	0.0
90480	SYRUP,CANE	26.00	269	0.00	0.00	0.86	73.14	0.0	73.2	13.0	...	0.0	0.000	0.000	0.000	0.0	NaN	NaN	NaN	NaN	0.0
90560	SNAIL,RAW	79.20	90	16.10	1.40	1.30	2.00	0.0	0.0	10.0	...	0.1	0.361	0.259	0.252	50.0	85.0	3 oz	NaN	NaN	0.0
93600	TURTLE,GREEN,RAW	78.50	89	19.80	0.50	1.20	0.00	0.0	0.0	118.0	...	0.1	0.127	0.088	0.170	50.0	85.0	3 oz	NaN	NaN	0.0

	age	weight	height	gender
2016-12-29 02:49:24.148388	23	75	176	male
2016-12-30 02:49:24.148388	23	75	176	male
2016-12-31 02:49:24.148388	23	75	176	male
2017-01-01 02:49:24.148388	23	75	176	male
2017-01-02 02:49:24.148388	23	75	176	male

	Shrt_Desc	Energ_Kcal	GmWt_1
No
1001	BUTTER,WITH SALT	717	227.00
1002	BUTTER,WHIPPED,WITH SALT	717	151.00
1003	BUTTER OIL,ANHYDROUS	876	205.00
1004	CHEESE,BLUE	353	28.35
1005	CHEESE,BRICK	371	132.00

	age	weight	height	gender	activity	kcal
2016-12-29 02:49:24.148388	23	75	176	male	very_heavy	3306.0
2016-12-30 02:49:24.148388	23	75	176	male	very_heavy	3306.0
2016-12-31 02:49:24.148388	23	75	176	male	heavy	3001.5
2017-01-01 02:49:24.148388	23	75	176	male	low	2088.0
2017-01-02 02:49:24.148388	23	75	176	male	very_heavy	3306.0

	Activity	Energy_Expenditure
0	Racquetball (recreational)	0.07
1	Kayaking (leisure)	0.04
2	Dancing (general)	0.08
3	Golf (walking + bag)	0.09
4	Running (5 mph, 12 min/mile)	0.12

	Unnamed: 0	age	weight	height	gender	activity	kcal	breakfast	breakfast_kcal	lunch	lunch_kcal	dinner	dinner_kcal	total_kcal	workout	workout_kcal	label
10	2017-01-08 02:24:13.995011	23	75	176	male	mid	2697.0	BEEF,CHUCK,CLOD,TOP BLADE,STEAK,LN & FAT,0" FA...	176	PUDDINGS,BANANA,DRY MIX,INST,W/ ADDED OIL	386	PEANUTS,ALL TYPES,RAW	567	1129	Weeding/Gardening	315.0	0
11	2017-01-09 02:24:13.995011	23	75	176	male	very_heavy	3306.0	CARROTS,CND,NO SALT,SOL&LIQUIDS	23	PACE,SALSA REFRIED BNS	60	FRANKFURTER,BF,UNHTD	313	396	Running 5.5 mph (11 min/mile)	630.0	1
12	2017-01-10 02:24:13.995011	23	75	176	male	heavy	3001.5	BABYFOOD,TEETHING BISCUITS	391	SESAME FLOUR,HIGH-FAT	526	KELLOGG'S EGGO LOWFAT BLUEBERRY NUTRI - GRAIN ...	208	1125	Carrying Groceries (light)	315.0	0
13	2017-01-11 02:24:13.995011	23	75	176	male	low	2088.0	PORK,FRSH,SPARERIBS,LN&FAT,CKD,BRSD	397	SAUCE,BARBECUE,SWT BABY RAY'S,ORIGINAL	192	KRAFT VELVEETA LT RED FAT PAST PROCESS CHS PRO...	222	811	Chopping Wood	405.0	0
14	2017-01-12 02:24:13.995011	23	75	176	male	very_heavy	3306.0	BEEF,TOP SIRLOIN,STEAK,LN & FAT,1/8" FAT,ALL G...	201	OIL,INDUSTRIAL,SOY,REFINED,FOR WOKS & LT FRYING	884	BABYFOOD,DINNER,VEG&CHICK,JR	53	1138	Sexual Activity (kissing, hugs)	90.0	0