-
Notifications
You must be signed in to change notification settings - Fork 0
/
results.txt
50 lines (43 loc) · 3.74 KB
/
results.txt
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
Results of SimpleNeuralNetwork at MNIST dataset.
Using 450 for train, 50 for validation, 100 for test:
NN [784, 30, 10]; LR: 0,000100; momentum: 0,000; epochs: 30; batch size: 10. acc = 10,000 %
NN [784, 30, 10]; LR: 0,000300; momentum: 0,000; epochs: 30; batch size: 10. acc = 23,000 %
NN [784, 30, 10]; LR: 0,001000; momentum: 0,000; epochs: 30; batch size: 10. acc = 15,000 %
NN [784, 30, 10]; LR: 0,003000; momentum: 0,000; epochs: 30; batch size: 10. acc = 27,000 %
NN [784, 30, 10]; LR: 0,010000; momentum: 0,000; epochs: 30; batch size: 10. acc = 48,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 56,000 %
NN [784, 30, 10]; LR: 0,100000; momentum: 0,000; epochs: 30; batch size: 10. acc = 47,000 %
NN [784, 30, 10]; LR: 0,300000; momentum: 0,000; epochs: 30; batch size: 10. acc = 25,000 %
NN [784, 30, 10]; LR: 1,000000; momentum: 0,000; epochs: 30; batch size: 10. acc = 19,000 %
Using 450 for train, 50 for validation, 100 for test:
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 1. acc = 22,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 3. acc = 32,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 52,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 50. acc = 39,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 100. acc = 22,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 500. acc = 9,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 1000. acc = 13,000 %
Using 450 for train, 50 for validation, 100 for test:
NN [784, 10, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 31,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 30. acc = 47,000 %
NN [784, 50, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 50. acc = 42,000 %
NN [784, 70, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 70. acc = 36,000 %
NN [784, 100, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 100. acc = 38,000 %
NN [784, 200, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 200. acc = 25,000 %
NN [784, 300, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 300. acc = 28,000 %
NN [784, 500, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 500. acc = 8,000 %
Using 450 for train, 50 for validation, 100 for test:
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 59,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,200; epochs: 30; batch size: 10. acc = 49,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,500; epochs: 30; batch size: 10. acc = 39,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,700; epochs: 30; batch size: 10. acc = 37,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,900; epochs: 30; batch size: 10. acc = 25,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,950; epochs: 30; batch size: 10. acc = 19,000 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,990; epochs: 30; batch size: 10. acc = 15,000 %
Using 4950 for train, 100 for validation, 500 for test:
NN [784, 10, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 53,600 %
NN [784, 30, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 66,800 %
NN [784, 100, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 79,600 %
NN [784, 300, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 82,600 %
Using 450 for train, 1000 for validation, 0 for test:
NN [784, 1000, 10]; LR: 0,030000; momentum: 0,000; epochs: 30; batch size: 10. acc = 76,300 % at 26