kaggle_leaf_classification/results.txt at master · MrChocolateMoose/kaggle_leaf_classification · GitHub

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
Best: 0.928283 using {'neurons': 192}
0.907071 (0.013627) with: {'neurons': 128}
0.928283 (0.012453) with: {'neurons': 192}
0.924242 (0.014845) with: {'neurons': 256}

===
192 neurons
loss: 0.0585 - acc: 0.9828 ==> KAGGLE SCORE 0.19379
loss: 0.0880 - acc: 0.9729 - val_loss: 14.6194 - val_acc: 0.0061
===

===
192 neruons, additional dropout in seperate models
loss: 0.1225 - acc: 0.9586 ==> KAGGLE SCORE: 0.19991
===

===
192 neruons, no dropout
loss: 0.0044 - acc: 0.9990 ==> KAGGLE SCORE: 0.37978
===


Best: 0.981818 using {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 384, 'joint_model_neurons': 384}
0.977778 (0.009367) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 192, 'joint_model_neurons': 384}
0.976768 (0.011780) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 256, 'joint_model_neurons': 384}
0.977778 (0.010400) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 384, 'joint_model_neurons': 384}
0.979798 (0.008451) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 192, 'joint_model_neurons': 512}
0.981818 (0.006851) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 256, 'joint_model_neurons': 512}
0.979798 (0.008451) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 384, 'joint_model_neurons': 512}
0.978788 (0.011693) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 192, 'joint_model_neurons': 768}
0.979798 (0.011517) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 256, 'joint_model_neurons': 768}
0.975758 (0.008081) with: {'dropout_prob': 0.20000000000000001, 'split_model_neurons': 384, 'joint_model_neurons': 768}
0.968687 (0.008689) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 192, 'joint_model_neurons': 384}
0.974747 (0.007142) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 256, 'joint_model_neurons': 384}
0.981818 (0.006851) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 384, 'joint_model_neurons': 384}
0.970707 (0.008689) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 192, 'joint_model_neurons': 512}
0.976768 (0.010400) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 256, 'joint_model_neurons': 512}
0.977778 (0.006851) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 384, 'joint_model_neurons': 512}
0.969697 (0.008451) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 192, 'joint_model_neurons': 768}
0.974747 (0.013170) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 256, 'joint_model_neurons': 768}
0.976768 (0.009367) with: {'dropout_prob': 0.30000000000000004, 'split_model_neurons': 384, 'joint_model_neurons': 768}
0.952525 (0.013401) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 192, 'joint_model_neurons': 384}
0.961616 (0.012205) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 256, 'joint_model_neurons': 384}
0.970707 (0.006700) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 384, 'joint_model_neurons': 384}
0.960606 (0.011248) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 192, 'joint_model_neurons': 512}
0.964646 (0.014285) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 256, 'joint_model_neurons': 512}
0.969697 (0.009583) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 384, 'joint_model_neurons': 512}
0.962626 (0.007559) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 192, 'joint_model_neurons': 768}
0.957576 (0.013014) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 256, 'joint_model_neurons': 768}
0.965657 (0.015118) with: {'dropout_prob': 0.40000000000000002, 'split_model_neurons': 384, 'joint_model_neurons': 768}