Newer
Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
import numpy as np
from .module import Module
from .sequential import Sequential
from .parallel import Parallel
from .optim import Optimizer
from .loss import LossFunction
from .scheduler import Scheduler
from data.dataLoader import DataLoader
from utility.progressbar import Progressbar
from .layer import Layer
from importlib import import_module
from inspect import signature
class Network(object):
"""
this class uses modules together with loss functions
and optimizers to handel the full neural network stack
"""
def __init__(self):
self.name = self.__class__.__name__
self.module = Sequential()
self.lossFunc = None
self.optim = None
self.scheduler = None
@property
def qualifiedName(self) -> tuple:
return self.__class__.__module__, self.__class__.__name__
def setComponent(self, component: Optimizer | LossFunction | Scheduler) -> None:
if isinstance(component, Optimizer):
self.optim = component
elif isinstance(component, LossFunction):
self.lossFunc = component
elif isinstance(component, Scheduler):
self.scheduler = component
else:
raise TypeError("The given component is not a valid type")
def append(self, layer: Layer, mode: str = 's') -> None:
if mode == 's':
if isinstance(self.module[-1], Parallel) and len(self.module[-1]) <= 1:
toPop = self.module[-1][0]
self.module.pop(-1)
self.module.append(toPop)
self.module.append(layer)
if mode == 'p':
if isinstance(self.module[-1], Parallel):
self.module[-1].append(layer)
else:
self.module.append(Parallel())
self.module[-1].append(layer)
def __call__(self, input: np.ndarray) -> np.ndarray:
"""
this makes using this class more convenient
"""
return self.forward(input)
def forward(self, input: np.ndarray) -> np.ndarray:
self.module.forward(input)
def backward(self, gradient: np.ndarray) -> np.ndarray:
self.module.backward(gradient)
def trainMode(self) -> None:
"""
sets every layer in the module into train mode
"""
self.module.train()
def evalMode(self) -> None:
"""
sets every layer in the module into eval mode
"""
self.module.eval()
def __str__(self):
printString = ""
printString += self.optim.name + ": " + str(self.optim.learningRate) + "\n"
if self.scheduler is not None:
printString += self.scheduler.name + "\n"
printString += self.lossFunc.name + "\n"
printString += str(self.module)
def train(self, data: DataLoader, epochs: int) -> None:
metrics = NetworkObservables(epochs)
epochs = epochs
for i in range(epochs):
data.trainMode()
self.trainMode()
length = len(data)
bar = Progressbar(f'epoch {str(i+1).zfill(len(str(epochs)))}/{epochs}', length, 55)
losses = []
for item in data:
inputs = item['data']
labels = item['labels']
prediction = network(inputs)
losses.append(self.loss(prediction, labels))
gradient = self.loss.backward()
self.optim.step(gradient)
bar.step()
data.evalMode()
self.evalMode()
accuracies = []
valLosses = []
for item in data:
inputs = item['data']
labels = item['labels']
prediction = network(inputs)
valLosses.append(self.loss(prediction, labels))
accuracies.append(np.sum(prediction.argmax(1) == labels.argmax(1)) / len(prediction))
bar.step()
metrics.update('losses', np.mean(losses))
metrics.update('validation', np.mean(valLosses))
metrics.update('accuracy', np.mean(accuracies))
metrics.update('learningRate', self.optim.learningRate)
metrics.print()
metrics.step()
data.fold()
if self.scheduler is not None:
self.scheduler.step()
def eval(self, data: DataLoader) -> np.ndarray:
self.evalEval()
length = len(data)
bar = Progressbar('evaluation', length)
predictions = []
for item in data:
inputs = item['data']
labels = item['labels']
predictions.append(network(inputs))
bar.step()
return np.concatnate(predictions)
def __getitem__(self, index):
return self.module[index]