spaceship_titanic🛸
Original Notebook : https://www.kaggle.com/code/parthavjoshi/spaceship-titanic
import os
import numpy as np
import pandas as pd
import gc
import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import Dataset, DataLoader
from sklearn.preprocessing import OneHotEncoder, LabelEncoder, StandardScaler
from sklearn.compose import ColumnTransformer
from sklearn.model_selection import train_test_split
from sklearn.model_selection import StratifiedKFold, cross_val_score
from sklearn.metrics import accuracy_score, roc_curve,auc, confusion_matrix,precision_recall_curve,precision_recall_curve
import tqdm
import warnings
warnings.simplefilter('ignore')
Load data
train = pd.read_csv(os.path.join(os.getcwd(), "data", "train.csv"))
train.head(2)
#> PassengerId HomePlanet CryoSleep ... VRDeck Name Transported
#> 0 0001_01 Europa False ... 0.0 Maham Ofracculy False
#> 1 0002_01 Earth False ... 44.0 Juanna Vines True
#>
#> [2 rows x 14 columns]
train.shape
#> (8693, 14)
test = pd.read_csv(os.path.join(os.getcwd(), "data", "test.csv"))
test.head(2)
#> PassengerId HomePlanet CryoSleep ... Spa VRDeck Name
#> 0 0013_01 Earth True ... 0.0 0.0 Nelly Carsoning
#> 1 0018_01 Earth False ... 2823.0 0.0 Lerome Peckers
#>
#> [2 rows x 13 columns]
test.shape
#> (4277, 13)
Exploring the data
train.describe()
#> Age RoomService ... Spa VRDeck
#> count 8514.000000 8512.000000 ... 8510.000000 8505.000000
#> mean 28.827930 224.687617 ... 311.138778 304.854791
#> std 14.489021 666.717663 ... 1136.705535 1145.717189
#> min 0.000000 0.000000 ... 0.000000 0.000000
#> 25% 19.000000 0.000000 ... 0.000000 0.000000
#> 50% 27.000000 0.000000 ... 0.000000 0.000000
#> 75% 38.000000 47.000000 ... 59.000000 46.000000
#> max 79.000000 14327.000000 ... 22408.000000 24133.000000
#>
#> [8 rows x 6 columns]
train.dtypes
#> PassengerId object
#> HomePlanet object
#> CryoSleep object
#> Cabin object
#> Destination object
#> Age float64
#> VIP object
#> RoomService float64
#> FoodCourt float64
#> ShoppingMall float64
#> Spa float64
#> VRDeck float64
#> Name object
#> Transported bool
#> dtype: object
train.nunique()
#> PassengerId 8693
#> HomePlanet 3
#> CryoSleep 2
#> Cabin 6560
#> Destination 3
#> Age 80
#> VIP 2
#> RoomService 1273
#> FoodCourt 1507
#> ShoppingMall 1115
#> Spa 1327
#> VRDeck 1306
#> Name 8473
#> Transported 2
#> dtype: int64
train.isna().sum()
#> PassengerId 0
#> HomePlanet 201
#> CryoSleep 217
#> Cabin 199
#> Destination 182
#> Age 179
#> VIP 203
#> RoomService 181
#> FoodCourt 183
#> ShoppingMall 208
#> Spa 183
#> VRDeck 188
#> Name 200
#> Transported 0
#> dtype: int64
test.isna().sum()
#> PassengerId 0
#> HomePlanet 87
#> CryoSleep 93
#> Cabin 100
#> Destination 92
#> Age 91
#> VIP 93
#> RoomService 82
#> FoodCourt 106
#> ShoppingMall 98
#> Spa 101
#> VRDeck 80
#> Name 94
#> dtype: int64
Data Preprocessing
train['is_train'] = True
test['is_train'] = False
data = pd.concat([train, test])
data.head()
#> PassengerId HomePlanet CryoSleep ... Name Transported is_train
#> 0 0001_01 Europa False ... Maham Ofracculy False True
#> 1 0002_01 Earth False ... Juanna Vines True True
#> 2 0003_01 Europa False ... Altark Susent False True
#> 3 0003_02 Europa False ... Solam Susent False True
#> 4 0004_01 Earth False ... Willy Santantines True True
#>
#> [5 rows x 15 columns]
def fill_missing_value(data, cols):
for c in cols:
data[c].fillna(data[c].median(skipna = True), inplace = True)
fill_missing_value(data, ['Age', 'RoomService', 'FoodCourt', 'ShoppingMall', 'Spa', 'VRDeck'])
data['HomePlanet'].fillna('Z', inplace = True)
def label_encode(data, col):
data[col] = data[col].astype(str)
data[col] = LabelEncoder().fit_transform(data[col])
return data[col]
data['HomePlanet'] = label_encode(data, 'HomePlanet')
data['CryoSleep'] = label_encode(data, 'CryoSleep')
data['VIP'] = label_encode(data, 'VIP')
data['Destination'] = label_encode(data, 'Destination')
mask = data['is_train'] == True
mask
#> 0 True
#> 1 True
#> 2 True
#> 3 True
#> 4 True
#> ...
#> 4272 False
#> 4273 False
#> 4274 False
#> 4275 False
#> 4276 False
#> Name: is_train, Length: 12970, dtype: bool
train = data[mask]
train.shape
#> (8693, 15)
test = data[~mask]
test.shape
#> (4277, 15)
train_data = train.drop(['is_train'], axis = 1)
test_data = test.drop(['is_train'], axis = 1)
train_data.isna().sum()
#> PassengerId 0
#> HomePlanet 0
#> CryoSleep 0
#> Cabin 199
#> Destination 0
#> Age 0
#> VIP 0
#> RoomService 0
#> FoodCourt 0
#> ShoppingMall 0
#> Spa 0
#> VRDeck 0
#> Name 200
#> Transported 0
#> dtype: int64
test_data.isna().sum()
#> PassengerId 0
#> HomePlanet 0
#> CryoSleep 0
#> Cabin 100
#> Destination 0
#> Age 0
#> VIP 0
#> RoomService 0
#> FoodCourt 0
#> ShoppingMall 0
#> Spa 0
#> VRDeck 0
#> Name 94
#> Transported 4277
#> dtype: int64
Model Training
train_data = train_data.dropna()
train_data.drop(['PassengerId', 'Cabin', 'Name'], axis = 1, inplace = True)
test_data.drop(['PassengerId', 'Cabin', 'Name'], axis = 1, inplace = True)
train_data['Transported'] = train_data['Transported'].map({True : 1, False : 0})
class Config:
lr = 1e-4
nb_epochs = 100
train_bs = 64
valid_bs = 64
train_split = 0.8
k_folds = 5
device = "cuda"
train_loss_fn = nn.BCEWithLogitsLoss()
valid_loss_fn = nn.BCEWithLogitsLoss()
feature_names = [
"HomePlanet", "CryoSleep", "Destination", "Age", "VIP", "RoomService", "FoodCourt", "ShoppingMall", "Spa", "VRDeck"
]
target_name = "Transported"
class SpaceshipTitanicModel(nn.Module):
def __init__(self, input_size = None, output_size = None):
super().__init__()
self.input_size = 10 if not input_size else input_size
self.output_size = 1 if not output_size else output_size
# model architecture
self.fc1 = nn.Linear(self.input_size, 1024)
self.fc2 = nn.Linear(1024, 768)
self.fc3 = nn.Linear(768, 128)
self.fc4 = nn.Linear(128, self.output_size)
self.relu = nn.ReLU()
self.sig = nn.Sigmoid()
def forward(self, x):
out = self.fc1(x)
out = self.relu(out)
out = self.fc2(out)
out = self.relu(out)
out = self.fc3(out)
out = self.relu(out)
out = self.fc4(out)
out = self.sig(out)
return out
def binary_acc(y_pred, y_test):
y_pred = torch.round(torch.sigmoid(y_pred))
correct = (y_pred == y_test).sum().float()
acc = correct / y_test.shape[0]
acc = torch.round(acc * 100)
return acc
class SpaceshipTitanicData(Dataset):
def __init__(self, features, target, is_test = False):
self.features = features
self.target = target
self.is_test = is_test
def __getitem__(self, idx):
data = self.features.values[idx]
if self.is_test:
return torch.tensor(data, dtype = torch.float32)
else:
target = self.target.values[idx]
return torch.tensor(data, dtype = torch.float32), torch.tensor(target, dtype = torch.float32)
def __len__(self):
return len(self.features)
def train_model(model, train_loader, optimizer, loss_fn, device):
"""
Training Function
"""
print("Training............")
# breakpoint()
with HiddenPrints():
model.train()
global y
global z
running_loss = 0
all_targets = []
all_preds = []
with HiddenPrints():
prog_bar = tqdm.tqdm(train_loader, total = len(train_loader), disable=True)
for x, y in prog_bar:
x = x.to(device, torch.float32)
y = y.to(device, torch.float32)
z = model(x)
train_loss = loss_fn(z, y)
acc = binary_acc(z, y)
train_loss.backward()
optimizer.step()
optimizer.zero_grad()
running_loss = running_loss + train_loss
prog_bar.set_description(f"train loss : {train_loss.item():.2f}")
all_targets.append(y.detach().cpu().numpy())
all_preds.append(z.detach().cpu().numpy())
return all_targets, all_preds
def valid_fn(model, tvalid_loader, loss_fn, device):
"""
Validation function
"""
with HiddenPrints():
model.eval()
running_loss = 0
all_targets = []
all_preds = []
with HiddenPrints():
prog_bar = tqdm.tqdm(valid_loader, total = len(valid_loader), disable=True)
for x, y in prog_bar:
x = x.to(device, torch.float32)
y = y.to(device, torch.float32)
z = model(x)
valid_loss = loss_fn(z, y)
acc = binary_acc(z, y)
running_loss = running_loss + valid_loss
prog_bar.set_description(f"Validation loss{valid_loss.item():.2f}")
all_targets.append(y.detach().cpu().numpy())
all_preds.append(z.detach().cpu().numpy())
print(f"Validation Loss: {running_loss:.4f}")
print(f"Acc : {acc:.3f}")
return all_targets, all_preds
if __name__ == "__main__":
data = train_data.sample(frac = 1).reset_index(drop = True)
kfold = StratifiedKFold(n_splits = Config.k_folds, shuffle = True)
for fold, (train_ids, valid_ids) in enumerate(kfold.split(data.drop(['Transported'], axis = 1), data['Transported'])):
print(f"FOLD {fold}")
print("-"*20)
train_ = data.iloc[train_ids]
valid_ = data.iloc[valid_ids]
train_dataset = SpaceshipTitanicData(
features = train_.drop(['Transported'], axis = 1), target = train_[['Transported']]
)
valid_dataset = SpaceshipTitanicData(
features = valid_.drop(['Transported'], axis = 1), target = valid_[['Transported']]
)
train_loader = DataLoader(train_dataset, batch_size = 32, shuffle = True)
valid_loader = DataLoader(valid_dataset, batch_size = 32, shuffle = True)
# model = SpaceshipTitanicModel(None, None)
with HiddenPrints():
model = SpaceshipTitanicModel(10, output_size = 1)
model.to(Config.device)
criterion = nn.BCEWithLogitsLoss()
optimizer = torch.optim.Adam(model.parameters(), lr = Config.lr)
print("[INFO]: Starting training! \n")
for epoch in range(1, Config.nb_epochs + 2):
print(f"{'='*20} Epoch: {epoch}/{Config.nb_epochs + 1} {'='*20}")
# breakpoint()
_, _ = train_model(model, train_loader, optimizer, Config.train_loss_fn, device = Config.device)
val_targets, val_preds = valid_fn(model, valid_loader, Config.valid_loss_fn, device = Config.device)
filepath = f"./model/fold_{fold}_model.pth"
torch.save(model.state_dict(), filepath)
#> FOLD 0
#> --------------------
#> [INFO]: Starting training!
#>
#> ==================== Epoch: 1/101 ====================
#> Training............
#> Validation Loss: 32.9721
#> Acc : 64.000
#> ==================== Epoch: 2/101 ====================
#> Training............
#> Validation Loss: 32.9451
#> Acc : 68.000
#> ==================== Epoch: 3/101 ====================
#> Training............
#> Validation Loss: 32.7616
#> Acc : 61.000
#> ==================== Epoch: 4/101 ====================
#> Training............
#> Validation Loss: 33.0184
#> Acc : 46.000
#> ==================== Epoch: 5/101 ====================
#> Training............
#> Validation Loss: 32.9175
#> Acc : 61.000
#> ==================== Epoch: 6/101 ====================
#> Training............
#> Validation Loss: 32.9008
#> Acc : 75.000
#> ==================== Epoch: 7/101 ====================
#> Training............
#> Validation Loss: 32.7833
#> Acc : 68.000
#> ==================== Epoch: 8/101 ====================
#> Training............
#> Validation Loss: 32.6718
#> Acc : 64.000
#> ==================== Epoch: 9/101 ====================
#> Training............
#> Validation Loss: 32.7567
#> Acc : 68.000
#> ==================== Epoch: 10/101 ====================
#> Training............
#> Validation Loss: 32.8100
#> Acc : 61.000
#> ==================== Epoch: 11/101 ====================
#> Training............
#> Validation Loss: 32.6540
#> Acc : 86.000
#> ==================== Epoch: 12/101 ====================
#> Training............
#> Validation Loss: 32.6463
#> Acc : 68.000
#> ==================== Epoch: 13/101 ====================
#> Training............
#> Validation Loss: 32.6278
#> Acc : 79.000
#> ==================== Epoch: 14/101 ====================
#> Training............
#> Validation Loss: 32.6276
#> Acc : 96.000
#> ==================== Epoch: 15/101 ====================
#> Training............
#> Validation Loss: 32.6208
#> Acc : 68.000
#> ==================== Epoch: 16/101 ====================
#> Training............
#> Validation Loss: 32.6880
#> Acc : 61.000
#> ==================== Epoch: 17/101 ====================
#> Training............
#> Validation Loss: 32.6252
#> Acc : 71.000
#> ==================== Epoch: 18/101 ====================
#> Training............
#> Validation Loss: 32.6143
#> Acc : 75.000
#> ==================== Epoch: 19/101 ====================
#> Training............
#> Validation Loss: 32.7284
#> Acc : 86.000
#> ==================== Epoch: 20/101 ====================
#> Training............
#> Validation Loss: 32.6267
#> Acc : 79.000
#> ==================== Epoch: 21/101 ====================
#> Training............
#> Validation Loss: 32.6392
#> Acc : 71.000
#> ==================== Epoch: 22/101 ====================
#> Training............
#> Validation Loss: 32.6944
#> Acc : 71.000
#> ==================== Epoch: 23/101 ====================
#> Training............
#> Validation Loss: 32.6182
#> Acc : 75.000
#> ==================== Epoch: 24/101 ====================
#> Training............
#> Validation Loss: 32.6009
#> Acc : 86.000
#> ==================== Epoch: 25/101 ====================
#> Training............
#> Validation Loss: 32.6116
#> Acc : 71.000
#> ==================== Epoch: 26/101 ====================
#> Training............
#> Validation Loss: 32.5519
#> Acc : 82.000
#> ==================== Epoch: 27/101 ====================
#> Training............
#> Validation Loss: 32.7343
#> Acc : 54.000
#> ==================== Epoch: 28/101 ====================
#> Training............
#> Validation Loss: 32.6637
#> Acc : 71.000
#> ==================== Epoch: 29/101 ====================
#> Training............
#> Validation Loss: 32.6478
#> Acc : 68.000
#> ==================== Epoch: 30/101 ====================
#> Training............
#> Validation Loss: 32.6703
#> Acc : 86.000
#> ==================== Epoch: 31/101 ====================
#> Training............
#> Validation Loss: 32.5538
#> Acc : 68.000
#> ==================== Epoch: 32/101 ====================
#> Training............
#> Validation Loss: 32.5788
#> Acc : 54.000
#> ==================== Epoch: 33/101 ====================
#> Training............
#> Validation Loss: 32.6384
#> Acc : 71.000
#> ==================== Epoch: 34/101 ====================
#> Training............
#> Validation Loss: 32.5269
#> Acc : 71.000
#> ==================== Epoch: 35/101 ====================
#> Training............
#> Validation Loss: 32.6504
#> Acc : 71.000
#> ==================== Epoch: 36/101 ====================
#> Training............
#> Validation Loss: 32.6172
#> Acc : 79.000
#> ==================== Epoch: 37/101 ====================
#> Training............
#> Validation Loss: 32.5784
#> Acc : 61.000
#> ==================== Epoch: 38/101 ====================
#> Training............
#> Validation Loss: 32.6027
#> Acc : 75.000
#> ==================== Epoch: 39/101 ====================
#> Training............
#> Validation Loss: 32.6837
#> Acc : 68.000
#> ==================== Epoch: 40/101 ====================
#> Training............
#> Validation Loss: 32.6208
#> Acc : 68.000
#> ==================== Epoch: 41/101 ====================
#> Training............
#> Validation Loss: 32.6247
#> Acc : 64.000
#> ==================== Epoch: 42/101 ====================
#> Training............
#> Validation Loss: 32.5238
#> Acc : 75.000
#> ==================== Epoch: 43/101 ====================
#> Training............
#> Validation Loss: 32.4938
#> Acc : 75.000
#> ==================== Epoch: 44/101 ====================
#> Training............
#> Validation Loss: 32.5215
#> Acc : 61.000
#> ==================== Epoch: 45/101 ====================
#> Training............
#> Validation Loss: 32.6543
#> Acc : 64.000
#> ==================== Epoch: 46/101 ====================
#> Training............
#> Validation Loss: 32.5570
#> Acc : 82.000
#> ==================== Epoch: 47/101 ====================
#> Training............
#> Validation Loss: 32.5015
#> Acc : 36.000
#> ==================== Epoch: 48/101 ====================
#> Training............
#> Validation Loss: 32.6139
#> Acc : 79.000
#> ==================== Epoch: 49/101 ====================
#> Training............
#> Validation Loss: 32.5770
#> Acc : 79.000
#> ==================== Epoch: 50/101 ====================
#> Training............
#> Validation Loss: 32.4856
#> Acc : 79.000
#> ==================== Epoch: 51/101 ====================
#> Training............
#> Validation Loss: 32.6654
#> Acc : 64.000
#> ==================== Epoch: 52/101 ====================
#> Training............
#> Validation Loss: 32.5877
#> Acc : 71.000
#> ==================== Epoch: 53/101 ====================
#> Training............
#> Validation Loss: 32.6560
#> Acc : 50.000
#> ==================== Epoch: 54/101 ====================
#> Training............
#> Validation Loss: 32.5226
#> Acc : 75.000
#> ==================== Epoch: 55/101 ====================
#> Training............
#> Validation Loss: 32.5638
#> Acc : 79.000
#> ==================== Epoch: 56/101 ====================
#> Training............
#> Validation Loss: 32.6685
#> Acc : 82.000
#> ==================== Epoch: 57/101 ====================
#> Training............
#> Validation Loss: 32.5034
#> Acc : 71.000
#> ==================== Epoch: 58/101 ====================
#> Training............
#> Validation Loss: 32.4747
#> Acc : 64.000
#> ==================== Epoch: 59/101 ====================
#> Training............
#> Validation Loss: 32.5946
#> Acc : 71.000
#> ==================== Epoch: 60/101 ====================
#> Training............
#> Validation Loss: 32.6357
#> Acc : 75.000
#> ==================== Epoch: 61/101 ====================
#> Training............
#> Validation Loss: 32.5137
#> Acc : 71.000
#> ==================== Epoch: 62/101 ====================
#> Training............
#> Validation Loss: 32.6340
#> Acc : 54.000
#> ==================== Epoch: 63/101 ====================
#> Training............
#> Validation Loss: 32.4892
#> Acc : 64.000
#> ==================== Epoch: 64/101 ====================
#> Training............
#> Validation Loss: 32.5161
#> Acc : 75.000
#> ==================== Epoch: 65/101 ====================
#> Training............
#> Validation Loss: 32.4710
#> Acc : 75.000
#> ==================== Epoch: 66/101 ====================
#> Training............
#> Validation Loss: 32.4833
#> Acc : 71.000
#> ==================== Epoch: 67/101 ====================
#> Training............
#> Validation Loss: 32.4834
#> Acc : 68.000
#> ==================== Epoch: 68/101 ====================
#> Training............
#> Validation Loss: 32.4528
#> Acc : 75.000
#> ==================== Epoch: 69/101 ====================
#> Training............
#> Validation Loss: 32.4887
#> Acc : 64.000
#> ==================== Epoch: 70/101 ====================
#> Training............
#> Validation Loss: 32.4838
#> Acc : 79.000
#> ==================== Epoch: 71/101 ====================
#> Training............
#> Validation Loss: 32.5077
#> Acc : 71.000
#> ==================== Epoch: 72/101 ====================
#> Training............
#> Validation Loss: 32.4879
#> Acc : 82.000
#> ==================== Epoch: 73/101 ====================
#> Training............
#> Validation Loss: 32.4515
#> Acc : 86.000
#> ==================== Epoch: 74/101 ====================
#> Training............
#> Validation Loss: 32.4931
#> Acc : 71.000
#> ==================== Epoch: 75/101 ====================
#> Training............
#> Validation Loss: 32.4663
#> Acc : 68.000
#> ==================== Epoch: 76/101 ====================
#> Training............
#> Validation Loss: 32.4320
#> Acc : 64.000
#> ==================== Epoch: 77/101 ====================
#> Training............
#> Validation Loss: 32.4683
#> Acc : 61.000
#> ==================== Epoch: 78/101 ====================
#> Training............
#> Validation Loss: 32.4915
#> Acc : 68.000
#> ==================== Epoch: 79/101 ====================
#> Training............
#> Validation Loss: 32.4562
#> Acc : 71.000
#> ==================== Epoch: 80/101 ====================
#> Training............
#> Validation Loss: 32.4750
#> Acc : 61.000
#> ==================== Epoch: 81/101 ====================
#> Training............
#> Validation Loss: 32.4483
#> Acc : 68.000
#> ==================== Epoch: 82/101 ====================
#> Training............
#> Validation Loss: 32.4506
#> Acc : 79.000
#> ==================== Epoch: 83/101 ====================
#> Training............
#> Validation Loss: 32.5008
#> Acc : 71.000
#> ==================== Epoch: 84/101 ====================
#> Training............
#> Validation Loss: 32.5901
#> Acc : 86.000
#> ==================== Epoch: 85/101 ====================
#> Training............
#> Validation Loss: 32.4524
#> Acc : 68.000
#> ==================== Epoch: 86/101 ====================
#> Training............
#> Validation Loss: 32.4343
#> Acc : 54.000
#> ==================== Epoch: 87/101 ====================
#> Training............
#> Validation Loss: 32.4879
#> Acc : 79.000
#> ==================== Epoch: 88/101 ====================
#> Training............
#> Validation Loss: 32.6224
#> Acc : 61.000
#> ==================== Epoch: 89/101 ====================
#> Training............
#> Validation Loss: 32.4471
#> Acc : 68.000
#> ==================== Epoch: 90/101 ====================
#> Training............
#> Validation Loss: 32.4686
#> Acc : 75.000
#> ==================== Epoch: 91/101 ====================
#> Training............
#> Validation Loss: 32.5360
#> Acc : 61.000
#> ==================== Epoch: 92/101 ====================
#> Training............
#> Validation Loss: 32.4304
#> Acc : 79.000
#> ==================== Epoch: 93/101 ====================
#> Training............
#> Validation Loss: 32.5232
#> Acc : 71.000
#> ==================== Epoch: 94/101 ====================
#> Training............
#> Validation Loss: 32.4515
#> Acc : 64.000
#> ==================== Epoch: 95/101 ====================
#> Training............
#> Validation Loss: 32.6347
#> Acc : 71.000
#> ==================== Epoch: 96/101 ====================
#> Training............
#> Validation Loss: 32.4506
#> Acc : 71.000
#> ==================== Epoch: 97/101 ====================
#> Training............
#> Validation Loss: 32.4264
#> Acc : 71.000
#> ==================== Epoch: 98/101 ====================
#> Training............
#> Validation Loss: 32.4248
#> Acc : 79.000
#> ==================== Epoch: 99/101 ====================
#> Training............
#> Validation Loss: 32.4205
#> Acc : 68.000
#> ==================== Epoch: 100/101 ====================
#> Training............
#> Validation Loss: 32.6022
#> Acc : 82.000
#> ==================== Epoch: 101/101 ====================
#> Training............
#> Validation Loss: 32.4567
#> Acc : 71.000
#> FOLD 1
#> --------------------
#> [INFO]: Starting training!
#>
#> ==================== Epoch: 1/101 ====================
#> Training............
#> Validation Loss: 32.0741
#> Acc : 74.000
#> ==================== Epoch: 2/101 ====================
#> Training............
#> Validation Loss: 32.0793
#> Acc : 78.000
#> ==================== Epoch: 3/101 ====================
#> Training............
#> Validation Loss: 32.0112
#> Acc : 67.000
#> ==================== Epoch: 4/101 ====================
#> Training............
#> Validation Loss: 31.9913
#> Acc : 70.000
#> ==================== Epoch: 5/101 ====================
#> Training............
#> Validation Loss: 31.9664
#> Acc : 74.000
#> ==================== Epoch: 6/101 ====================
#> Training............
#> Validation Loss: 31.9633
#> Acc : 67.000
#> ==================== Epoch: 7/101 ====================
#> Training............
#> Validation Loss: 32.0019
#> Acc : 59.000
#> ==================== Epoch: 8/101 ====================
#> Training............
#> Validation Loss: 31.9894
#> Acc : 78.000
#> ==================== Epoch: 9/101 ====================
#> Training............
#> Validation Loss: 31.9527
#> Acc : 56.000
#> ==================== Epoch: 10/101 ====================
#> Training............
#> Validation Loss: 42.0798
#> Acc : 52.000
#> ==================== Epoch: 11/101 ====================
#> Training............
#> Validation Loss: 42.0882
#> Acc : 44.000
#> ==================== Epoch: 12/101 ====================
#> Training............
#> Validation Loss: 42.0450
#> Acc : 63.000
#> ==================== Epoch: 13/101 ====================
#> Training............
#> Validation Loss: 33.1219
#> Acc : 70.000
#> ==================== Epoch: 14/101 ====================
#> Training............
#> Validation Loss: 32.6106
#> Acc : 89.000
#> ==================== Epoch: 15/101 ====================
#> Training............
#> Validation Loss: 31.8354
#> Acc : 74.000
#> ==================== Epoch: 16/101 ====================
#> Training............
#> Validation Loss: 32.1190
#> Acc : 81.000
#> ==================== Epoch: 17/101 ====================
#> Training............
#> Validation Loss: 32.2270
#> Acc : 70.000
#> ==================== Epoch: 18/101 ====================
#> Training............
#> Validation Loss: 32.4035
#> Acc : 67.000
#> ==================== Epoch: 19/101 ====================
#> Training............
#> Validation Loss: 31.9205
#> Acc : 59.000
#> ==================== Epoch: 20/101 ====================
#> Training............
#> Validation Loss: 31.7721
#> Acc : 93.000
#> ==================== Epoch: 21/101 ====================
#> Training............
#> Validation Loss: 31.7417
#> Acc : 93.000
#> ==================== Epoch: 22/101 ====================
#> Training............
#> Validation Loss: 32.2490
#> Acc : 81.000
#> ==================== Epoch: 23/101 ====================
#> Training............
#> Validation Loss: 31.7279
#> Acc : 85.000
#> ==================== Epoch: 24/101 ====================
#> Training............
#> Validation Loss: 31.5132
#> Acc : 78.000
#> ==================== Epoch: 25/101 ====================
#> Training............
#> Validation Loss: 31.8979
#> Acc : 74.000
#> ==================== Epoch: 26/101 ====================
#> Training............
#> Validation Loss: 31.9311
#> Acc : 78.000
#> ==================== Epoch: 27/101 ====================
#> Training............
#> Validation Loss: 31.9809
#> Acc : 74.000
#> ==================== Epoch: 28/101 ====================
#> Training............
#> Validation Loss: 31.9282
#> Acc : 78.000
#> ==================== Epoch: 29/101 ====================
#> Training............
#> Validation Loss: 32.0917
#> Acc : 81.000
#> ==================== Epoch: 30/101 ====================
#> Training............
#> Validation Loss: 31.9609
#> Acc : 56.000
#> ==================== Epoch: 31/101 ====================
#> Training............
#> Validation Loss: 31.9452
#> Acc : 93.000
#> ==================== Epoch: 32/101 ====================
#> Training............
#> Validation Loss: 32.0786
#> Acc : 78.000
#> ==================== Epoch: 33/101 ====================
#> Training............
#> Validation Loss: 31.9342
#> Acc : 70.000
#> ==================== Epoch: 34/101 ====================
#> Training............
#> Validation Loss: 31.9449
#> Acc : 81.000
#> ==================== Epoch: 35/101 ====================
#> Training............
#> Validation Loss: 31.9726
#> Acc : 70.000
#> ==================== Epoch: 36/101 ====================
#> Training............
#> Validation Loss: 31.9385
#> Acc : 67.000
#> ==================== Epoch: 37/101 ====================
#> Training............
#> Validation Loss: 31.9732
#> Acc : 81.000
#> ==================== Epoch: 38/101 ====================
#> Training............
#> Validation Loss: 31.9723
#> Acc : 78.000
#> ==================== Epoch: 39/101 ====================
#> Training............
#> Validation Loss: 31.9287
#> Acc : 78.000
#> ==================== Epoch: 40/101 ====================
#> Training............
#> Validation Loss: 31.9447
#> Acc : 78.000
#> ==================== Epoch: 41/101 ====================
#> Training............
#> Validation Loss: 32.0138
#> Acc : 67.000
#> ==================== Epoch: 42/101 ====================
#> Training............
#> Validation Loss: 31.9472
#> Acc : 74.000
#> ==================== Epoch: 43/101 ====================
#> Training............
#> Validation Loss: 32.0060
#> Acc : 81.000
#> ==================== Epoch: 44/101 ====================
#> Training............
#> Validation Loss: 31.9748
#> Acc : 81.000
#> ==================== Epoch: 45/101 ====================
#> Training............
#> Validation Loss: 31.9444
#> Acc : 81.000
#> ==================== Epoch: 46/101 ====================
#> Training............
#> Validation Loss: 32.2896
#> Acc : 78.000
#> ==================== Epoch: 47/101 ====================
#> Training............
#> Validation Loss: 32.3560
#> Acc : 74.000
#> ==================== Epoch: 48/101 ====================
#> Training............
#> Validation Loss: 32.1566
#> Acc : 74.000
#> ==================== Epoch: 49/101 ====================
#> Training............
#> Validation Loss: 31.9675
#> Acc : 63.000
#> ==================== Epoch: 50/101 ====================
#> Training............
#> Validation Loss: 31.9837
#> Acc : 67.000
#> ==================== Epoch: 51/101 ====================
#> Training............
#> Validation Loss: 31.9219
#> Acc : 81.000
#> ==================== Epoch: 52/101 ====================
#> Training............
#> Validation Loss: 31.9377
#> Acc : 81.000
#> ==================== Epoch: 53/101 ====================
#> Training............
#> Validation Loss: 31.9479
#> Acc : 78.000
#> ==================== Epoch: 54/101 ====================
#> Training............
#> Validation Loss: 31.9445
#> Acc : 70.000
#> ==================== Epoch: 55/101 ====================
#> Training............
#> Validation Loss: 32.0322
#> Acc : 78.000
#> ==================== Epoch: 56/101 ====================
#> Training............
#> Validation Loss: 31.9841
#> Acc : 63.000
#> ==================== Epoch: 57/101 ====================
#> Training............
#> Validation Loss: 32.0148
#> Acc : 89.000
#> ==================== Epoch: 58/101 ====================
#> Training............
#> Validation Loss: 32.0192
#> Acc : 78.000
#> ==================== Epoch: 59/101 ====================
#> Training............
#> Validation Loss: 32.0356
#> Acc : 81.000
#> ==================== Epoch: 60/101 ====================
#> Training............
#> Validation Loss: 31.9601
#> Acc : 85.000
#> ==================== Epoch: 61/101 ====================
#> Training............
#> Validation Loss: 31.9739
#> Acc : 78.000
#> ==================== Epoch: 62/101 ====================
#> Training............
#> Validation Loss: 31.9333
#> Acc : 89.000
#> ==================== Epoch: 63/101 ====================
#> Training............
#> Validation Loss: 31.9509
#> Acc : 67.000
#> ==================== Epoch: 64/101 ====================
#> Training............
#> Validation Loss: 31.9794
#> Acc : 74.000
#> ==================== Epoch: 65/101 ====================
#> Training............
#> Validation Loss: 31.9404
#> Acc : 67.000
#> ==================== Epoch: 66/101 ====================
#> Training............
#> Validation Loss: 31.9598
#> Acc : 59.000
#> ==================== Epoch: 67/101 ====================
#> Training............
#> Validation Loss: 31.9947
#> Acc : 59.000
#> ==================== Epoch: 68/101 ====================
#> Training............
#> Validation Loss: 31.9563
#> Acc : 67.000
#> ==================== Epoch: 69/101 ====================
#> Training............
#> Validation Loss: 31.9281
#> Acc : 78.000
#> ==================== Epoch: 70/101 ====================
#> Training............
#> Validation Loss: 32.1664
#> Acc : 81.000
#> ==================== Epoch: 71/101 ====================
#> Training............
#> Validation Loss: 31.9624
#> Acc : 56.000
#> ==================== Epoch: 72/101 ====================
#> Training............
#> Validation Loss: 31.9584
#> Acc : 70.000
#> ==================== Epoch: 73/101 ====================
#> Training............
#> Validation Loss: 31.9461
#> Acc : 81.000
#> ==================== Epoch: 74/101 ====================
#> Training............
#> Validation Loss: 31.9632
#> Acc : 78.000
#> ==================== Epoch: 75/101 ====================
#> Training............
#> Validation Loss: 31.9631
#> Acc : 78.000
#> ==================== Epoch: 76/101 ====================
#> Training............
#> Validation Loss: 31.9898
#> Acc : 70.000
#> ==================== Epoch: 77/101 ====================
#> Training............
#> Validation Loss: 31.9634
#> Acc : 67.000
#> ==================== Epoch: 78/101 ====================
#> Training............
#> Validation Loss: 31.9527
#> Acc : 89.000
#> ==================== Epoch: 79/101 ====================
#> Training............
#> Validation Loss: 32.0380
#> Acc : 67.000
#> ==================== Epoch: 80/101 ====================
#> Training............
#> Validation Loss: 31.9615
#> Acc : 70.000
#> ==================== Epoch: 81/101 ====================
#> Training............
#> Validation Loss: 31.9719
#> Acc : 78.000
#> ==================== Epoch: 82/101 ====================
#> Training............
#> Validation Loss: 31.9661
#> Acc : 70.000
#> ==================== Epoch: 83/101 ====================
#> Training............
#> Validation Loss: 31.9625
#> Acc : 74.000
#> ==================== Epoch: 84/101 ====================
#> Training............
#> Validation Loss: 31.9785
#> Acc : 78.000
#> ==================== Epoch: 85/101 ====================
#> Training............
#> Validation Loss: 31.9491
#> Acc : 70.000
#> ==================== Epoch: 86/101 ====================
#> Training............
#> Validation Loss: 31.9629
#> Acc : 70.000
#> ==================== Epoch: 87/101 ====================
#> Training............
#> Validation Loss: 31.9587
#> Acc : 81.000
#> ==================== Epoch: 88/101 ====================
#> Training............
#> Validation Loss: 32.1117
#> Acc : 70.000
#> ==================== Epoch: 89/101 ====================
#> Training............
#> Validation Loss: 31.9552
#> Acc : 59.000
#> ==================== Epoch: 90/101 ====================
#> Training............
#> Validation Loss: 31.9508
#> Acc : 74.000
#> ==================== Epoch: 91/101 ====================
#> Training............
#> Validation Loss: 31.9446
#> Acc : 70.000
#> ==================== Epoch: 92/101 ====================
#> Training............
#> Validation Loss: 31.9456
#> Acc : 89.000
#> ==================== Epoch: 93/101 ====================
#> Training............
#> Validation Loss: 31.9771
#> Acc : 44.000
#> ==================== Epoch: 94/101 ====================
#> Training............
#> Validation Loss: 31.9220
#> Acc : 74.000
#> ==================== Epoch: 95/101 ====================
#> Training............
#> Validation Loss: 31.9390
#> Acc : 78.000
#> ==================== Epoch: 96/101 ====================
#> Training............
#> Validation Loss: 31.9546
#> Acc : 63.000
#> ==================== Epoch: 97/101 ====================
#> Training............
#> Validation Loss: 31.9418
#> Acc : 74.000
#> ==================== Epoch: 98/101 ====================
#> Training............
#> Validation Loss: 31.9513
#> Acc : 74.000
#> ==================== Epoch: 99/101 ====================
#> Training............
#> Validation Loss: 31.9636
#> Acc : 81.000
#> ==================== Epoch: 100/101 ====================
#> Training............
#> Validation Loss: 31.9668
#> Acc : 70.000
#> ==================== Epoch: 101/101 ====================
#> Training............
#> Validation Loss: 31.9640
#> Acc : 81.000
#> FOLD 2
#> --------------------
#> [INFO]: Starting training!
#>
#> ==================== Epoch: 1/101 ====================
#> Training............
#> Validation Loss: 32.7515
#> Acc : 81.000
#> ==================== Epoch: 2/101 ====================
#> Training............
#> Validation Loss: 32.5457
#> Acc : 74.000
#> ==================== Epoch: 3/101 ====================
#> Training............
#> Validation Loss: 33.1386
#> Acc : 70.000
#> ==================== Epoch: 4/101 ====================
#> Training............
#> Validation Loss: 32.9074
#> Acc : 74.000
#> ==================== Epoch: 5/101 ====================
#> Training............
#> Validation Loss: 32.3112
#> Acc : 59.000
#> ==================== Epoch: 6/101 ====================
#> Training............
#> Validation Loss: 32.2465
#> Acc : 81.000
#> ==================== Epoch: 7/101 ====================
#> Training............
#> Validation Loss: 32.2952
#> Acc : 70.000
#> ==================== Epoch: 8/101 ====================
#> Training............
#> Validation Loss: 32.2554
#> Acc : 59.000
#> ==================== Epoch: 9/101 ====================
#> Training............
#> Validation Loss: 32.2470
#> Acc : 93.000
#> ==================== Epoch: 10/101 ====================
#> Training............
#> Validation Loss: 32.2554
#> Acc : 63.000
#> ==================== Epoch: 11/101 ====================
#> Training............
#> Validation Loss: 32.2492
#> Acc : 78.000
#> ==================== Epoch: 12/101 ====================
#> Training............
#> Validation Loss: 32.2571
#> Acc : 78.000
#> ==================== Epoch: 13/101 ====================
#> Training............
#> Validation Loss: 32.2611
#> Acc : 78.000
#> ==================== Epoch: 14/101 ====================
#> Training............
#> Validation Loss: 32.2393
#> Acc : 63.000
#> ==================== Epoch: 15/101 ====================
#> Training............
#> Validation Loss: 32.2470
#> Acc : 85.000
#> ==================== Epoch: 16/101 ====================
#> Training............
#> Validation Loss: 32.2468
#> Acc : 67.000
#> ==================== Epoch: 17/101 ====================
#> Training............
#> Validation Loss: 32.2287
#> Acc : 67.000
#> ==================== Epoch: 18/101 ====================
#> Training............
#> Validation Loss: 32.2556
#> Acc : 63.000
#> ==================== Epoch: 19/101 ====================
#> Training............
#> Validation Loss: 32.2520
#> Acc : 78.000
#> ==================== Epoch: 20/101 ====================
#> Training............
#> Validation Loss: 32.2724
#> Acc : 78.000
#> ==================== Epoch: 21/101 ====================
#> Training............
#> Validation Loss: 32.9061
#> Acc : 74.000
#> ==================== Epoch: 22/101 ====================
#> Training............
#> Validation Loss: 32.2217
#> Acc : 74.000
#> ==================== Epoch: 23/101 ====================
#> Training............
#> Validation Loss: 32.2754
#> Acc : 63.000
#> ==================== Epoch: 24/101 ====================
#> Training............
#> Validation Loss: 32.2897
#> Acc : 56.000
#> ==================== Epoch: 25/101 ====================
#> Training............
#> Validation Loss: 32.2237
#> Acc : 78.000
#> ==================== Epoch: 26/101 ====================
#> Training............
#> Validation Loss: 32.3220
#> Acc : 70.000
#> ==================== Epoch: 27/101 ====================
#> Training............
#> Validation Loss: 32.2558
#> Acc : 56.000
#> ==================== Epoch: 28/101 ====================
#> Training............
#> Validation Loss: 32.2462
#> Acc : 59.000
#> ==================== Epoch: 29/101 ====================
#> Training............
#> Validation Loss: 32.3054
#> Acc : 70.000
#> ==================== Epoch: 30/101 ====================
#> Training............
#> Validation Loss: 32.2338
#> Acc : 78.000
#> ==================== Epoch: 31/101 ====================
#> Training............
#> Validation Loss: 32.2343
#> Acc : 63.000
#> ==================== Epoch: 32/101 ====================
#> Training............
#> Validation Loss: 32.2325
#> Acc : 74.000
#> ==================== Epoch: 33/101 ====================
#> Training............
#> Validation Loss: 32.2339
#> Acc : 67.000
#> ==================== Epoch: 34/101 ====================
#> Training............
#> Validation Loss: 32.2460
#> Acc : 81.000
#> ==================== Epoch: 35/101 ====================
#> Training............
#> Validation Loss: 32.2630
#> Acc : 74.000
#> ==================== Epoch: 36/101 ====================
#> Training............
#> Validation Loss: 32.2373
#> Acc : 59.000
#> ==================== Epoch: 37/101 ====================
#> Training............
#> Validation Loss: 32.2395
#> Acc : 59.000
#> ==================== Epoch: 38/101 ====================
#> Training............
#> Validation Loss: 32.2701
#> Acc : 74.000
#> ==================== Epoch: 39/101 ====================
#> Training............
#> Validation Loss: 32.3130
#> Acc : 81.000
#> ==================== Epoch: 40/101 ====================
#> Training............
#> Validation Loss: 32.2522
#> Acc : 67.000
#> ==================== Epoch: 41/101 ====================
#> Training............
#> Validation Loss: 32.2397
#> Acc : 63.000
#> ==================== Epoch: 42/101 ====================
#> Training............
#> Validation Loss: 32.3460
#> Acc : 70.000
#> ==================== Epoch: 43/101 ====================
#> Training............
#> Validation Loss: 32.9451
#> Acc : 81.000
#> ==================== Epoch: 44/101 ====================
#> Training............
#> Validation Loss: 32.3240
#> Acc : 70.000
#> ==================== Epoch: 45/101 ====================
#> Training............
#> Validation Loss: 32.4281
#> Acc : 67.000
#> ==================== Epoch: 46/101 ====================
#> Training............
#> Validation Loss: 32.2544
#> Acc : 70.000
#> ==================== Epoch: 47/101 ====================
#> Training............
#> Validation Loss: 32.2397
#> Acc : 81.000
#> ==================== Epoch: 48/101 ====================
#> Training............
#> Validation Loss: 32.2309
#> Acc : 81.000
#> ==================== Epoch: 49/101 ====================
#> Training............
#> Validation Loss: 32.2377
#> Acc : 74.000
#> ==================== Epoch: 50/101 ====================
#> Training............
#> Validation Loss: 32.2374
#> Acc : 74.000
#> ==================== Epoch: 51/101 ====================
#> Training............
#> Validation Loss: 32.2315
#> Acc : 74.000
#> ==================== Epoch: 52/101 ====================
#> Training............
#> Validation Loss: 32.2288
#> Acc : 85.000
#> ==================== Epoch: 53/101 ====================
#> Training............
#> Validation Loss: 32.2446
#> Acc : 67.000
#> ==================== Epoch: 54/101 ====================
#> Training............
#> Validation Loss: 32.2045
#> Acc : 85.000
#> ==================== Epoch: 55/101 ====================
#> Training............
#> Validation Loss: 32.2060
#> Acc : 89.000
#> ==================== Epoch: 56/101 ====================
#> Training............
#> Validation Loss: 32.4252
#> Acc : 74.000
#> ==================== Epoch: 57/101 ====================
#> Training............
#> Validation Loss: 32.2831
#> Acc : 70.000
#> ==================== Epoch: 58/101 ====================
#> Training............
#> Validation Loss: 32.1882
#> Acc : 89.000
#> ==================== Epoch: 59/101 ====================
#> Training............
#> Validation Loss: 32.2554
#> Acc : 67.000
#> ==================== Epoch: 60/101 ====================
#> Training............
#> Validation Loss: 32.1626
#> Acc : 74.000
#> ==================== Epoch: 61/101 ====================
#> Training............
#> Validation Loss: 32.2494
#> Acc : 63.000
#> ==================== Epoch: 62/101 ====================
#> Training............
#> Validation Loss: 32.1637
#> Acc : 78.000
#> ==================== Epoch: 63/101 ====================
#> Training............
#> Validation Loss: 32.3259
#> Acc : 74.000
#> ==================== Epoch: 64/101 ====================
#> Training............
#> Validation Loss: 32.2530
#> Acc : 63.000
#> ==================== Epoch: 65/101 ====================
#> Training............
#> Validation Loss: 32.2392
#> Acc : 70.000
#> ==================== Epoch: 66/101 ====================
#> Training............
#> Validation Loss: 32.2068
#> Acc : 81.000
#> ==================== Epoch: 67/101 ====================
#> Training............
#> Validation Loss: 32.2431
#> Acc : 74.000
#> ==================== Epoch: 68/101 ====================
#> Training............
#> Validation Loss: 32.6418
#> Acc : 89.000
#> ==================== Epoch: 69/101 ====================
#> Training............
#> Validation Loss: 32.2659
#> Acc : 63.000
#> ==================== Epoch: 70/101 ====================
#> Training............
#> Validation Loss: 32.2025
#> Acc : 70.000
#> ==================== Epoch: 71/101 ====================
#> Training............
#> Validation Loss: 32.2328
#> Acc : 78.000
#> ==================== Epoch: 72/101 ====================
#> Training............
#> Validation Loss: 32.1969
#> Acc : 78.000
#> ==================== Epoch: 73/101 ====================
#> Training............
#> Validation Loss: 32.1521
#> Acc : 81.000
#> ==================== Epoch: 74/101 ====================
#> Training............
#> Validation Loss: 32.1717
#> Acc : 67.000
#> ==================== Epoch: 75/101 ====================
#> Training............
#> Validation Loss: 32.2132
#> Acc : 70.000
#> ==================== Epoch: 76/101 ====================
#> Training............
#> Validation Loss: 32.2570
#> Acc : 70.000
#> ==================== Epoch: 77/101 ====================
#> Training............
#> Validation Loss: 32.2470
#> Acc : 78.000
#> ==================== Epoch: 78/101 ====================
#> Training............
#> Validation Loss: 32.2170
#> Acc : 81.000
#> ==================== Epoch: 79/101 ====================
#> Training............
#> Validation Loss: 32.1842
#> Acc : 74.000
#> ==================== Epoch: 80/101 ====================
#> Training............
#> Validation Loss: 32.2305
#> Acc : 81.000
#> ==================== Epoch: 81/101 ====================
#> Training............
#> Validation Loss: 32.2114
#> Acc : 85.000
#> ==================== Epoch: 82/101 ====================
#> Training............
#> Validation Loss: 32.1777
#> Acc : 63.000
#> ==================== Epoch: 83/101 ====================
#> Training............
#> Validation Loss: 32.2724
#> Acc : 78.000
#> ==================== Epoch: 84/101 ====================
#> Training............
#> Validation Loss: 32.1935
#> Acc : 70.000
#> ==================== Epoch: 85/101 ====================
#> Training............
#> Validation Loss: 32.1724
#> Acc : 74.000
#> ==================== Epoch: 86/101 ====================
#> Training............
#> Validation Loss: 32.2101
#> Acc : 85.000
#> ==================== Epoch: 87/101 ====================
#> Training............
#> Validation Loss: 32.2215
#> Acc : 78.000
#> ==================== Epoch: 88/101 ====================
#> Training............
#> Validation Loss: 32.2357
#> Acc : 63.000
#> ==================== Epoch: 89/101 ====================
#> Training............
#> Validation Loss: 32.2769
#> Acc : 81.000
#> ==================== Epoch: 90/101 ====================
#> Training............
#> Validation Loss: 32.2098
#> Acc : 70.000
#> ==================== Epoch: 91/101 ====================
#> Training............
#> Validation Loss: 32.2341
#> Acc : 70.000
#> ==================== Epoch: 92/101 ====================
#> Training............
#> Validation Loss: 32.2384
#> Acc : 74.000
#> ==================== Epoch: 93/101 ====================
#> Training............
#> Validation Loss: 32.1831
#> Acc : 81.000
#> ==================== Epoch: 94/101 ====================
#> Training............
#> Validation Loss: 32.2071
#> Acc : 85.000
#> ==================== Epoch: 95/101 ====================
#> Training............
#> Validation Loss: 32.2244
#> Acc : 74.000
#> ==================== Epoch: 96/101 ====================
#> Training............
#> Validation Loss: 32.2809
#> Acc : 56.000
#> ==================== Epoch: 97/101 ====================
#> Training............
#> Validation Loss: 32.1799
#> Acc : 67.000
#> ==================== Epoch: 98/101 ====================
#> Training............
#> Validation Loss: 32.1987
#> Acc : 74.000
#> ==================== Epoch: 99/101 ====================
#> Training............
#> Validation Loss: 32.2270
#> Acc : 74.000
#> ==================== Epoch: 100/101 ====================
#> Training............
#> Validation Loss: 32.2151
#> Acc : 85.000
#> ==================== Epoch: 101/101 ====================
#> Training............
#> Validation Loss: 32.2073
#> Acc : 81.000
#> FOLD 3
#> --------------------
#> [INFO]: Starting training!
#>
#> ==================== Epoch: 1/101 ====================
#> Training............
#> Validation Loss: 32.4459
#> Acc : 89.000
#> ==================== Epoch: 2/101 ====================
#> Training............
#> Validation Loss: 32.4004
#> Acc : 74.000
#> ==================== Epoch: 3/101 ====================
#> Training............
#> Validation Loss: 32.3176
#> Acc : 78.000
#> ==================== Epoch: 4/101 ====================
#> Training............
#> Validation Loss: 32.2679
#> Acc : 59.000
#> ==================== Epoch: 5/101 ====================
#> Training............
#> Validation Loss: 32.2924
#> Acc : 78.000
#> ==================== Epoch: 6/101 ====================
#> Training............
#> Validation Loss: 32.3137
#> Acc : 70.000
#> ==================== Epoch: 7/101 ====================
#> Training............
#> Validation Loss: 32.2433
#> Acc : 74.000
#> ==================== Epoch: 8/101 ====================
#> Training............
#> Validation Loss: 32.2696
#> Acc : 70.000
#> ==================== Epoch: 9/101 ====================
#> Training............
#> Validation Loss: 32.1701
#> Acc : 78.000
#> ==================== Epoch: 10/101 ====================
#> Training............
#> Validation Loss: 32.1451
#> Acc : 70.000
#> ==================== Epoch: 11/101 ====================
#> Training............
#> Validation Loss: 32.2696
#> Acc : 74.000
#> ==================== Epoch: 12/101 ====================
#> Training............
#> Validation Loss: 32.2576
#> Acc : 78.000
#> ==================== Epoch: 13/101 ====================
#> Training............
#> Validation Loss: 32.2771
#> Acc : 81.000
#> ==================== Epoch: 14/101 ====================
#> Training............
#> Validation Loss: 32.1397
#> Acc : 78.000
#> ==================== Epoch: 15/101 ====================
#> Training............
#> Validation Loss: 32.1111
#> Acc : 63.000
#> ==================== Epoch: 16/101 ====================
#> Training............
#> Validation Loss: 32.2222
#> Acc : 70.000
#> ==================== Epoch: 17/101 ====================
#> Training............
#> Validation Loss: 32.1745
#> Acc : 63.000
#> ==================== Epoch: 18/101 ====================
#> Training............
#> Validation Loss: 32.2462
#> Acc : 74.000
#> ==================== Epoch: 19/101 ====================
#> Training............
#> Validation Loss: 32.1151
#> Acc : 67.000
#> ==================== Epoch: 20/101 ====================
#> Training............
#> Validation Loss: 32.1163
#> Acc : 74.000
#> ==================== Epoch: 21/101 ====================
#> Training............
#> Validation Loss: 32.2448
#> Acc : 81.000
#> ==================== Epoch: 22/101 ====================
#> Training............
#> Validation Loss: 32.2887
#> Acc : 70.000
#> ==================== Epoch: 23/101 ====================
#> Training............
#> Validation Loss: 32.3035
#> Acc : 67.000
#> ==================== Epoch: 24/101 ====================
#> Training............
#> Validation Loss: 32.2374
#> Acc : 81.000
#> ==================== Epoch: 25/101 ====================
#> Training............
#> Validation Loss: 32.4458
#> Acc : 70.000
#> ==================== Epoch: 26/101 ====================
#> Training............
#> Validation Loss: 32.3264
#> Acc : 81.000
#> ==================== Epoch: 27/101 ====================
#> Training............
#> Validation Loss: 32.1449
#> Acc : 78.000
#> ==================== Epoch: 28/101 ====================
#> Training............
#> Validation Loss: 32.2199
#> Acc : 74.000
#> ==================== Epoch: 29/101 ====================
#> Training............
#> Validation Loss: 32.2986
#> Acc : 81.000
#> ==================== Epoch: 30/101 ====================
#> Training............
#> Validation Loss: 32.1877
#> Acc : 74.000
#> ==================== Epoch: 31/101 ====================
#> Training............
#> Validation Loss: 32.3110
#> Acc : 78.000
#> ==================== Epoch: 32/101 ====================
#> Training............
#> Validation Loss: 32.2457
#> Acc : 63.000
#> ==================== Epoch: 33/101 ====================
#> Training............
#> Validation Loss: 32.2906
#> Acc : 81.000
#> ==================== Epoch: 34/101 ====================
#> Training............
#> Validation Loss: 32.2804
#> Acc : 70.000
#> ==================== Epoch: 35/101 ====================
#> Training............
#> Validation Loss: 32.2962
#> Acc : 74.000
#> ==================== Epoch: 36/101 ====================
#> Training............
#> Validation Loss: 32.2793
#> Acc : 85.000
#> ==================== Epoch: 37/101 ====================
#> Training............
#> Validation Loss: 32.3082
#> Acc : 78.000
#> ==================== Epoch: 38/101 ====================
#> Training............
#> Validation Loss: 32.2681
#> Acc : 81.000
#> ==================== Epoch: 39/101 ====================
#> Training............
#> Validation Loss: 32.2872
#> Acc : 70.000
#> ==================== Epoch: 40/101 ====================
#> Training............
#> Validation Loss: 32.2857
#> Acc : 78.000
#> ==================== Epoch: 41/101 ====================
#> Training............
#> Validation Loss: 32.3191
#> Acc : 70.000
#> ==================== Epoch: 42/101 ====================
#> Training............
#> Validation Loss: 32.3137
#> Acc : 67.000
#> ==================== Epoch: 43/101 ====================
#> Training............
#> Validation Loss: 32.3099
#> Acc : 67.000
#> ==================== Epoch: 44/101 ====================
#> Training............
#> Validation Loss: 32.2905
#> Acc : 89.000
#> ==================== Epoch: 45/101 ====================
#> Training............
#> Validation Loss: 32.3119
#> Acc : 56.000
#> ==================== Epoch: 46/101 ====================
#> Training............
#> Validation Loss: 32.3027
#> Acc : 63.000
#> ==================== Epoch: 47/101 ====================
#> Training............
#> Validation Loss: 32.2989
#> Acc : 81.000
#> ==================== Epoch: 48/101 ====================
#> Training............
#> Validation Loss: 32.1966
#> Acc : 63.000
#> ==================== Epoch: 49/101 ====================
#> Training............
#> Validation Loss: 32.2982
#> Acc : 78.000
#> ==================== Epoch: 50/101 ====================
#> Training............
#> Validation Loss: 32.3127
#> Acc : 85.000
#> ==================== Epoch: 51/101 ====================
#> Training............
#> Validation Loss: 32.3110
#> Acc : 74.000
#> ==================== Epoch: 52/101 ====================
#> Training............
#> Validation Loss: 32.3562
#> Acc : 74.000
#> ==================== Epoch: 53/101 ====================
#> Training............
#> Validation Loss: 32.2441
#> Acc : 63.000
#> ==================== Epoch: 54/101 ====================
#> Training............
#> Validation Loss: 32.3322
#> Acc : 59.000
#> ==================== Epoch: 55/101 ====================
#> Training............
#> Validation Loss: 32.3293
#> Acc : 78.000
#> ==================== Epoch: 56/101 ====================
#> Training............
#> Validation Loss: 32.3212
#> Acc : 70.000
#> ==================== Epoch: 57/101 ====================
#> Training............
#> Validation Loss: 32.2967
#> Acc : 85.000
#> ==================== Epoch: 58/101 ====================
#> Training............
#> Validation Loss: 32.3245
#> Acc : 70.000
#> ==================== Epoch: 59/101 ====================
#> Training............
#> Validation Loss: 32.2777
#> Acc : 89.000
#> ==================== Epoch: 60/101 ====================
#> Training............
#> Validation Loss: 32.3346
#> Acc : 63.000
#> ==================== Epoch: 61/101 ====================
#> Training............
#> Validation Loss: 32.2933
#> Acc : 70.000
#> ==================== Epoch: 62/101 ====================
#> Training............
#> Validation Loss: 32.3149
#> Acc : 70.000
#> ==================== Epoch: 63/101 ====================
#> Training............
#> Validation Loss: 32.3489
#> Acc : 70.000
#> ==================== Epoch: 64/101 ====================
#> Training............
#> Validation Loss: 32.3209
#> Acc : 59.000
#> ==================== Epoch: 65/101 ====================
#> Training............
#> Validation Loss: 32.3399
#> Acc : 67.000
#> ==================== Epoch: 66/101 ====================
#> Training............
#> Validation Loss: 32.3396
#> Acc : 81.000
#> ==================== Epoch: 67/101 ====================
#> Training............
#> Validation Loss: 32.3359
#> Acc : 81.000
#> ==================== Epoch: 68/101 ====================
#> Training............
#> Validation Loss: 32.3114
#> Acc : 63.000
#> ==================== Epoch: 69/101 ====================
#> Training............
#> Validation Loss: 32.3262
#> Acc : 67.000
#> ==================== Epoch: 70/101 ====================
#> Training............
#> Validation Loss: 32.3311
#> Acc : 63.000
#> ==================== Epoch: 71/101 ====================
#> Training............
#> Validation Loss: 32.3334
#> Acc : 67.000
#> ==================== Epoch: 72/101 ====================
#> Training............
#> Validation Loss: 32.3105
#> Acc : 74.000
#> ==================== Epoch: 73/101 ====================
#> Training............
#> Validation Loss: 32.2733
#> Acc : 63.000
#> ==================== Epoch: 74/101 ====================
#> Training............
#> Validation Loss: 32.3717
#> Acc : 74.000
#> ==================== Epoch: 75/101 ====================
#> Training............
#> Validation Loss: 32.3453
#> Acc : 70.000
#> ==================== Epoch: 76/101 ====================
#> Training............
#> Validation Loss: 32.3088
#> Acc : 89.000
#> ==================== Epoch: 77/101 ====================
#> Training............
#> Validation Loss: 32.3270
#> Acc : 78.000
#> ==================== Epoch: 78/101 ====================
#> Training............
#> Validation Loss: 32.3451
#> Acc : 78.000
#> ==================== Epoch: 79/101 ====================
#> Training............
#> Validation Loss: 32.3111
#> Acc : 70.000
#> ==================== Epoch: 80/101 ====================
#> Training............
#> Validation Loss: 32.3300
#> Acc : 85.000
#> ==================== Epoch: 81/101 ====================
#> Training............
#> Validation Loss: 32.3218
#> Acc : 74.000
#> ==================== Epoch: 82/101 ====================
#> Training............
#> Validation Loss: 32.4535
#> Acc : 74.000
#> ==================== Epoch: 83/101 ====================
#> Training............
#> Validation Loss: 32.3112
#> Acc : 81.000
#> ==================== Epoch: 84/101 ====================
#> Training............
#> Validation Loss: 32.3338
#> Acc : 70.000
#> ==================== Epoch: 85/101 ====================
#> Training............
#> Validation Loss: 32.3199
#> Acc : 74.000
#> ==================== Epoch: 86/101 ====================
#> Training............
#> Validation Loss: 32.3067
#> Acc : 85.000
#> ==================== Epoch: 87/101 ====================
#> Training............
#> Validation Loss: 32.3575
#> Acc : 63.000
#> ==================== Epoch: 88/101 ====================
#> Training............
#> Validation Loss: 32.3516
#> Acc : 85.000
#> ==================== Epoch: 89/101 ====================
#> Training............
#> Validation Loss: 32.1795
#> Acc : 74.000
#> ==================== Epoch: 90/101 ====================
#> Training............
#> Validation Loss: 32.2965
#> Acc : 67.000
#> ==================== Epoch: 91/101 ====================
#> Training............
#> Validation Loss: 32.3653
#> Acc : 78.000
#> ==================== Epoch: 92/101 ====================
#> Training............
#> Validation Loss: 32.3756
#> Acc : 74.000
#> ==================== Epoch: 93/101 ====================
#> Training............
#> Validation Loss: 32.2672
#> Acc : 78.000
#> ==================== Epoch: 94/101 ====================
#> Training............
#> Validation Loss: 32.1775
#> Acc : 59.000
#> ==================== Epoch: 95/101 ====================
#> Training............
#> Validation Loss: 32.2946
#> Acc : 59.000
#> ==================== Epoch: 96/101 ====================
#> Training............
#> Validation Loss: 32.3335
#> Acc : 67.000
#> ==================== Epoch: 97/101 ====================
#> Training............
#> Validation Loss: 32.3237
#> Acc : 78.000
#> ==================== Epoch: 98/101 ====================
#> Training............
#> Validation Loss: 32.2961
#> Acc : 78.000
#> ==================== Epoch: 99/101 ====================
#> Training............
#> Validation Loss: 32.2880
#> Acc : 67.000
#> ==================== Epoch: 100/101 ====================
#> Training............
#> Validation Loss: 32.1370
#> Acc : 93.000
#> ==================== Epoch: 101/101 ====================
#> Training............
#> Validation Loss: 32.2427
#> Acc : 74.000
#> FOLD 4
#> --------------------
#> [INFO]: Starting training!
#>
#> ==================== Epoch: 1/101 ====================
#> Training............
#> Validation Loss: 33.2708
#> Acc : 63.000
#> ==================== Epoch: 2/101 ====================
#> Training............
#> Validation Loss: 32.3070
#> Acc : 70.000
#> ==================== Epoch: 3/101 ====================
#> Training............
#> Validation Loss: 32.2484
#> Acc : 70.000
#> ==================== Epoch: 4/101 ====================
#> Training............
#> Validation Loss: 32.2537
#> Acc : 74.000
#> ==================== Epoch: 5/101 ====================
#> Training............
#> Validation Loss: 32.2348
#> Acc : 70.000
#> ==================== Epoch: 6/101 ====================
#> Training............
#> Validation Loss: 32.1730
#> Acc : 78.000
#> ==================== Epoch: 7/101 ====================
#> Training............
#> Validation Loss: 32.1084
#> Acc : 63.000
#> ==================== Epoch: 8/101 ====================
#> Training............
#> Validation Loss: 32.1946
#> Acc : 81.000
#> ==================== Epoch: 9/101 ====================
#> Training............
#> Validation Loss: 32.1127
#> Acc : 70.000
#> ==================== Epoch: 10/101 ====================
#> Training............
#> Validation Loss: 32.1157
#> Acc : 67.000
#> ==================== Epoch: 11/101 ====================
#> Training............
#> Validation Loss: 32.1735
#> Acc : 70.000
#> ==================== Epoch: 12/101 ====================
#> Training............
#> Validation Loss: 32.0610
#> Acc : 59.000
#> ==================== Epoch: 13/101 ====================
#> Training............
#> Validation Loss: 32.1058
#> Acc : 85.000
#> ==================== Epoch: 14/101 ====================
#> Training............
#> Validation Loss: 32.2133
#> Acc : 63.000
#> ==================== Epoch: 15/101 ====================
#> Training............
#> Validation Loss: 32.1410
#> Acc : 74.000
#> ==================== Epoch: 16/101 ====================
#> Training............
#> Validation Loss: 32.1577
#> Acc : 70.000
#> ==================== Epoch: 17/101 ====================
#> Training............
#> Validation Loss: 32.1245
#> Acc : 89.000
#> ==================== Epoch: 18/101 ====================
#> Training............
#> Validation Loss: 32.1331
#> Acc : 74.000
#> ==================== Epoch: 19/101 ====================
#> Training............
#> Validation Loss: 32.0699
#> Acc : 74.000
#> ==================== Epoch: 20/101 ====================
#> Training............
#> Validation Loss: 32.1260
#> Acc : 70.000
#> ==================== Epoch: 21/101 ====================
#> Training............
#> Validation Loss: 32.0995
#> Acc : 67.000
#> ==================== Epoch: 22/101 ====================
#> Training............
#> Validation Loss: 32.2034
#> Acc : 85.000
#> ==================== Epoch: 23/101 ====================
#> Training............
#> Validation Loss: 32.1913
#> Acc : 56.000
#> ==================== Epoch: 24/101 ====================
#> Training............
#> Validation Loss: 32.1626
#> Acc : 74.000
#> ==================== Epoch: 25/101 ====================
#> Training............
#> Validation Loss: 32.1170
#> Acc : 93.000
#> ==================== Epoch: 26/101 ====================
#> Training............
#> Validation Loss: 32.1352
#> Acc : 74.000
#> ==================== Epoch: 27/101 ====================
#> Training............
#> Validation Loss: 31.9213
#> Acc : 63.000
#> ==================== Epoch: 28/101 ====================
#> Training............
#> Validation Loss: 31.9532
#> Acc : 70.000
#> ==================== Epoch: 29/101 ====================
#> Training............
#> Validation Loss: 32.1377
#> Acc : 67.000
#> ==================== Epoch: 30/101 ====================
#> Training............
#> Validation Loss: 32.1315
#> Acc : 67.000
#> ==================== Epoch: 31/101 ====================
#> Training............
#> Validation Loss: 32.1613
#> Acc : 74.000
#> ==================== Epoch: 32/101 ====================
#> Training............
#> Validation Loss: 32.1136
#> Acc : 85.000
#> ==================== Epoch: 33/101 ====================
#> Training............
#> Validation Loss: 32.1270
#> Acc : 74.000
#> ==================== Epoch: 34/101 ====================
#> Training............
#> Validation Loss: 32.1646
#> Acc : 56.000
#> ==================== Epoch: 35/101 ====================
#> Training............
#> Validation Loss: 32.1336
#> Acc : 70.000
#> ==================== Epoch: 36/101 ====================
#> Training............
#> Validation Loss: 32.1141
#> Acc : 78.000
#> ==================== Epoch: 37/101 ====================
#> Training............
#> Validation Loss: 32.1370
#> Acc : 85.000
#> ==================== Epoch: 38/101 ====================
#> Training............
#> Validation Loss: 32.1512
#> Acc : 74.000
#> ==================== Epoch: 39/101 ====================
#> Training............
#> Validation Loss: 32.1385
#> Acc : 74.000
#> ==================== Epoch: 40/101 ====================
#> Training............
#> Validation Loss: 32.0590
#> Acc : 81.000
#> ==================== Epoch: 41/101 ====================
#> Training............
#> Validation Loss: 32.1215
#> Acc : 67.000
#> ==================== Epoch: 42/101 ====================
#> Training............
#> Validation Loss: 32.1529
#> Acc : 78.000
#> ==================== Epoch: 43/101 ====================
#> Training............
#> Validation Loss: 32.1561
#> Acc : 89.000
#> ==================== Epoch: 44/101 ====================
#> Training............
#> Validation Loss: 32.1800
#> Acc : 74.000
#> ==================== Epoch: 45/101 ====================
#> Training............
#> Validation Loss: 32.0987
#> Acc : 81.000
#> ==================== Epoch: 46/101 ====================
#> Training............
#> Validation Loss: 32.0948
#> Acc : 96.000
#> ==================== Epoch: 47/101 ====================
#> Training............
#> Validation Loss: 31.9811
#> Acc : 78.000
#> ==================== Epoch: 48/101 ====================
#> Training............
#> Validation Loss: 32.1245
#> Acc : 70.000
#> ==================== Epoch: 49/101 ====================
#> Training............
#> Validation Loss: 32.1620
#> Acc : 63.000
#> ==================== Epoch: 50/101 ====================
#> Training............
#> Validation Loss: 32.1584
#> Acc : 70.000
#> ==================== Epoch: 51/101 ====================
#> Training............
#> Validation Loss: 32.1604
#> Acc : 74.000
#> ==================== Epoch: 52/101 ====================
#> Training............
#> Validation Loss: 32.1582
#> Acc : 74.000
#> ==================== Epoch: 53/101 ====================
#> Training............
#> Validation Loss: 32.1735
#> Acc : 78.000
#> ==================== Epoch: 54/101 ====================
#> Training............
#> Validation Loss: 32.1184
#> Acc : 74.000
#> ==================== Epoch: 55/101 ====================
#> Training............
#> Validation Loss: 32.1632
#> Acc : 63.000
#> ==================== Epoch: 56/101 ====================
#> Training............
#> Validation Loss: 31.9969
#> Acc : 67.000
#> ==================== Epoch: 57/101 ====================
#> Training............
#> Validation Loss: 32.1353
#> Acc : 70.000
#> ==================== Epoch: 58/101 ====================
#> Training............
#> Validation Loss: 31.9599
#> Acc : 70.000
#> ==================== Epoch: 59/101 ====================
#> Training............
#> Validation Loss: 31.9234
#> Acc : 81.000
#> ==================== Epoch: 60/101 ====================
#> Training............
#> Validation Loss: 32.0813
#> Acc : 78.000
#> ==================== Epoch: 61/101 ====================
#> Training............
#> Validation Loss: 32.1591
#> Acc : 56.000
#> ==================== Epoch: 62/101 ====================
#> Training............
#> Validation Loss: 32.1370
#> Acc : 63.000
#> ==================== Epoch: 63/101 ====================
#> Training............
#> Validation Loss: 31.8901
#> Acc : 74.000
#> ==================== Epoch: 64/101 ====================
#> Training............
#> Validation Loss: 31.9577
#> Acc : 70.000
#> ==================== Epoch: 65/101 ====================
#> Training............
#> Validation Loss: 31.8985
#> Acc : 70.000
#> ==================== Epoch: 66/101 ====================
#> Training............
#> Validation Loss: 31.9028
#> Acc : 63.000
#> ==================== Epoch: 67/101 ====================
#> Training............
#> Validation Loss: 31.9257
#> Acc : 56.000
#> ==================== Epoch: 68/101 ====================
#> Training............
#> Validation Loss: 32.1350
#> Acc : 81.000
#> ==================== Epoch: 69/101 ====================
#> Training............
#> Validation Loss: 32.1109
#> Acc : 81.000
#> ==================== Epoch: 70/101 ====================
#> Training............
#> Validation Loss: 31.9052
#> Acc : 74.000
#> ==================== Epoch: 71/101 ====================
#> Training............
#> Validation Loss: 31.9997
#> Acc : 85.000
#> ==================== Epoch: 72/101 ====================
#> Training............
#> Validation Loss: 31.9319
#> Acc : 78.000
#> ==================== Epoch: 73/101 ====================
#> Training............
#> Validation Loss: 31.9095
#> Acc : 78.000
#> ==================== Epoch: 74/101 ====================
#> Training............
#> Validation Loss: 31.9159
#> Acc : 67.000
#> ==================== Epoch: 75/101 ====================
#> Training............
#> Validation Loss: 31.9633
#> Acc : 85.000
#> ==================== Epoch: 76/101 ====================
#> Training............
#> Validation Loss: 32.0942
#> Acc : 74.000
#> ==================== Epoch: 77/101 ====================
#> Training............
#> Validation Loss: 32.1012
#> Acc : 74.000
#> ==================== Epoch: 78/101 ====================
#> Training............
#> Validation Loss: 32.0188
#> Acc : 78.000
#> ==================== Epoch: 79/101 ====================
#> Training............
#> Validation Loss: 32.1427
#> Acc : 78.000
#> ==================== Epoch: 80/101 ====================
#> Training............
#> Validation Loss: 31.9381
#> Acc : 70.000
#> ==================== Epoch: 81/101 ====================
#> Training............
#> Validation Loss: 32.1456
#> Acc : 74.000
#> ==================== Epoch: 82/101 ====================
#> Training............
#> Validation Loss: 32.0446
#> Acc : 74.000
#> ==================== Epoch: 83/101 ====================
#> Training............
#> Validation Loss: 31.9292
#> Acc : 89.000
#> ==================== Epoch: 84/101 ====================
#> Training............
#> Validation Loss: 31.9324
#> Acc : 81.000
#> ==================== Epoch: 85/101 ====================
#> Training............
#> Validation Loss: 31.9263
#> Acc : 67.000
#> ==================== Epoch: 86/101 ====================
#> Training............
#> Validation Loss: 32.0984
#> Acc : 59.000
#> ==================== Epoch: 87/101 ====================
#> Training............
#> Validation Loss: 32.0853
#> Acc : 74.000
#> ==================== Epoch: 88/101 ====================
#> Training............
#> Validation Loss: 32.1416
#> Acc : 67.000
#> ==================== Epoch: 89/101 ====================
#> Training............
#> Validation Loss: 31.9688
#> Acc : 70.000
#> ==================== Epoch: 90/101 ====================
#> Training............
#> Validation Loss: 32.0012
#> Acc : 85.000
#> ==================== Epoch: 91/101 ====================
#> Training............
#> Validation Loss: 31.9488
#> Acc : 70.000
#> ==================== Epoch: 92/101 ====================
#> Training............
#> Validation Loss: 31.9313
#> Acc : 67.000
#> ==================== Epoch: 93/101 ====================
#> Training............
#> Validation Loss: 31.9252
#> Acc : 74.000
#> ==================== Epoch: 94/101 ====================
#> Training............
#> Validation Loss: 31.9020
#> Acc : 70.000
#> ==================== Epoch: 95/101 ====================
#> Training............
#> Validation Loss: 31.9364
#> Acc : 81.000
#> ==================== Epoch: 96/101 ====================
#> Training............
#> Validation Loss: 31.9293
#> Acc : 85.000
#> ==================== Epoch: 97/101 ====================
#> Training............
#> Validation Loss: 31.9496
#> Acc : 70.000
#> ==================== Epoch: 98/101 ====================
#> Training............
#> Validation Loss: 31.9407
#> Acc : 78.000
#> ==================== Epoch: 99/101 ====================
#> Training............
#> Validation Loss: 31.9400
#> Acc : 81.000
#> ==================== Epoch: 100/101 ====================
#> Training............
#> Validation Loss: 31.9351
#> Acc : 56.000
#> ==================== Epoch: 101/101 ====================
#> Training............
#> Validation Loss: 31.9966
#> Acc : 67.000
@torch.no_grad()
def inference(model, states_list, test_dataloader, device):
"""
Do inference for different model folds
"""
with HiddenPrints():
model.eval()
all_preds = []
for state in states_list:
print(f"State: {state}")
state_dict = torch.load(state)
model.load_state_dict(state_dict)
model = model.to(device)
# Clean
del state_dict
gc.collect()
preds = []
prog = tqdm.tqdm(test_dataloader, total = len(test_dataloader), disable=True)
for x in prog:
x = x.to(device, dtype = torch.float32)
outputs = model(x)
preds.append(outputs.squeeze(-1).cpu().detach().numpy())
all_preds.append(np.concatenate(preds))
# Clean
gc.collect()
torch.cuda.empty_cache()
return all_preds
model_dir = "./model"
states_list = [os.path.join(model_dir, x) for x in os.listdir(model_dir) if x.endswith(".pth")]
test_dataset = SpaceshipTitanicData(features = test_data.drop(['Transported'], axis = 1),
target = None,
is_test = True
)
test_loader = DataLoader(test_dataset, batch_size = 32, shuffle = False)
print("Predictions for all folds")
#> Predictions for all folds
predictions = inference(model, states_list, test_loader, Config.device)
#> State: ./model/fold_1_model.pth
#> State: ./model/fold_4_model.pth
#> State: ./model/fold_3_model.pth
#> State: ./model/fold_2_model.pth
#> State: ./model/fold_0_model.pth
pred = pd.DataFrame(predictions).T.mean(axis = 1).tolist()
pred = [True if p >= 0.5 else False for p in pred]
submission = pd.DataFrame({
"PassengerId" : test['PassengerId'],
"Transported" : pred
})
submission.to_csv(os.path.join(os.getcwd(), "model", "submission.csv"), index = False)