Skip to content

Instantly share code, notes, and snippets.

@bamos
Created June 13, 2016 18:01
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save bamos/a5102f890f16a391084a0950618ef477 to your computer and use it in GitHub Desktop.
Save bamos/a5102f890f16a391084a0950618ef477 to your computer and use it in GitHub Desktop.
nan
1 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> (7) -> (8) -> (9) -> (10) -> (11) -> (12) -> (13) -> (14) -> (15) -> (16) -> (17) -> (18) -> (19) -> (20) -> (21) -> (22) -> (23) -> (24) -> output]
(1): nn.SpatialConvolution(3 -> 64, 7x7, 2,2, 3,3)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialMaxPooling(3x3, 2,2, 1,1)
(5): nn.SpatialCrossMapLRN
(6): nn.SpatialConvolution(64 -> 64, 1x1)
(7): nn.SpatialBatchNormalization
(8): nn.ReLU
(9): nn.SpatialConvolution(64 -> 192, 3x3, 1,1, 1,1)
(10): nn.SpatialBatchNormalization
(11): nn.ReLU
(12): nn.SpatialCrossMapLRN
(13): nn.SpatialMaxPooling(3x3, 2,2, 1,1)
(14): nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(192 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(192 -> 16, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(16 -> 32, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.SpatialMaxPooling(3x3, 2,2)
| (2): nn.SpatialConvolution(192 -> 32, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(192 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
(15): nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(256 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(256 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(256 -> 64, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(256 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
(16): nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(320 -> 128, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(128 -> 256, 3x3, 2,2, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(320 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 2,2, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
}
... -> output
}
(17): nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 192, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(640 -> 128, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(640 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
(18): nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 160, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(160 -> 256, 3x3, 2,2, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 64, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(64 -> 128, 5x5, 2,2, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
}
... -> output
}
(19): nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(1024 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(1024 -> 96, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(1024 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
(20): nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(736 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.SpatialMaxPooling(3x3, 2,2)
| (2): nn.SpatialConvolution(736 -> 96, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(736 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
(21): nn.SpatialAveragePooling(3x3, 1,1)
(22): nn.View(736)
(23): nn.Linear(736 -> 128)
(24): nn.Normalize(2)
}
nan
2 nn.SpatialConvolution(3 -> 64, 7x7, 2,2, 3,3)
28428.669774754
3 nn.SpatialBatchNormalization
47485.344581918
4 nn.ReLU
47485.344581918
5 nn.SpatialMaxPooling(3x3, 2,2, 1,1)
29610.480826637
6 nn.SpatialCrossMapLRN
29607.20285791
7 nn.SpatialConvolution(64 -> 64, 1x1)
-46215.649390483
8 nn.SpatialBatchNormalization
5586.9869854718
9 nn.ReLU
5586.9869854718
10 nn.SpatialConvolution(64 -> 192, 3x3, 1,1, 1,1)
-243945.2438823
11 nn.SpatialBatchNormalization
14310.947079424
12 nn.ReLU
14310.947079424
13 nn.SpatialCrossMapLRN
14310.764782931
14 nn.SpatialMaxPooling(3x3, 2,2, 1,1)
9299.4612118299
15 nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(192 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(192 -> 16, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(16 -> 32, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.SpatialMaxPooling(3x3, 2,2)
| (2): nn.SpatialConvolution(192 -> 32, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(192 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
7372.8237418521
16 nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(192 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(192 -> 16, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(16 -> 32, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.SpatialMaxPooling(3x3, 2,2)
| (2): nn.SpatialConvolution(192 -> 32, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(192 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
7372.8237418521
17 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(192 -> 96, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
3041.5289890589
18 nn.SpatialConvolution(192 -> 96, 1x1)
-4221.4973005245
19 nn.SpatialBatchNormalization
3381.8238505099
20 nn.ReLU
3381.8238505099
21 nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
7723.8720341735
22 nn.SpatialBatchNormalization
3041.5289890589
23 nn.ReLU
3041.5289890589
24 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(192 -> 16, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(16 -> 32, 5x5, 1,1, 2,2)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
1097.3607423482
25 nn.SpatialConvolution(192 -> 16, 1x1)
-1095.6663571014
26 nn.SpatialBatchNormalization
533.20889493544
27 nn.ReLU
533.20889493544
28 nn.SpatialConvolution(16 -> 32, 5x5, 1,1, 2,2)
560.78297215328
29 nn.SpatialBatchNormalization
1097.3607423482
30 nn.ReLU
1097.3607423482
31 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
(2): nn.SpatialConvolution(192 -> 32, 1x1)
(3): nn.SpatialBatchNormalization
(4): nn.ReLU
}
255.2191820778
32 nn.SpatialMaxPooling(3x3, 2,2)
2537.3004810991
33 nn.SpatialConvolution(192 -> 32, 1x1)
260.92130632512
34 nn.SpatialBatchNormalization
255.2191820778
35 nn.ReLU
255.2191820778
36 nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(192 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
2978.7148283671
37 nn.SpatialConvolution(192 -> 64, 1x1)
-1925.0513102314
38 nn.SpatialBatchNormalization
2978.7148283671
39 nn.ReLU
2978.7148283671
40 nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(256 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(256 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(256 -> 64, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(256 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
4192.5248958138
41 nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(256 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(256 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(256 -> 64, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(256 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
4192.5248958138
42 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(256 -> 96, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
1740.6393912397
43 nn.SpatialConvolution(256 -> 96, 1x1)
-5744.7828996833
44 nn.SpatialBatchNormalization
2135.0005637809
45 nn.ReLU
2135.0005637809
46 nn.SpatialConvolution(96 -> 128, 3x3, 1,1, 1,1)
-31627.677911759
47 nn.SpatialBatchNormalization
1740.6393912397
48 nn.ReLU
1740.6393912397
49 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(256 -> 32, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
1100.3323725164
50 nn.SpatialConvolution(256 -> 32, 1x1)
-2077.8073705663
51 nn.SpatialBatchNormalization
816.2209900188
52 nn.ReLU
816.2209900188
53 nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
-9842.4670063686
54 nn.SpatialBatchNormalization
1100.3323725164
55 nn.ReLU
1100.3323725164
56 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Square
(2): nn.SpatialAveragePooling(3x3, 3,3)
(3): nn.MulConstant
(4): nn.Sqrt
}
(2): nn.SpatialConvolution(256 -> 64, 1x1)
(3): nn.SpatialBatchNormalization
(4): nn.ReLU
}
123.4834844321
57 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Square
(2): nn.SpatialAveragePooling(3x3, 3,3)
(3): nn.MulConstant
(4): nn.Sqrt
}
2995.9608779663
58 nn.Square
5307.9655835477
59 nn.SpatialAveragePooling(3x3, 3,3)
5307.9655962903
60 nn.MulConstant
5307.9655962903
61 nn.Sqrt
2995.9608779663
62 nn.SpatialConvolution(256 -> 64, 1x1)
-1013.5408580252
63 nn.SpatialBatchNormalization
123.4834844321
64 nn.ReLU
123.4834844321
65 nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(256 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
1228.0696476256
66 nn.SpatialConvolution(256 -> 64, 1x1)
-4882.411236471
67 nn.SpatialBatchNormalization
1228.0696476256
68 nn.ReLU
1228.0696476256
69 nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(320 -> 128, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(128 -> 256, 3x3, 2,2, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(320 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 2,2, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
}
... -> output
}
4118.1789596067
70 nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(320 -> 128, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(128 -> 256, 3x3, 2,2, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(320 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 2,2, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
}
... -> output
}
4118.1789596067
71 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(320 -> 128, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(128 -> 256, 3x3, 2,2, 1,1)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
1590.7774455175
72 nn.SpatialConvolution(320 -> 128, 1x1)
-6934.5462531344
73 nn.SpatialBatchNormalization
2216.6457054392
74 nn.ReLU
2216.6457054392
75 nn.SpatialConvolution(128 -> 256, 3x3, 2,2, 1,1)
-5044.9248673413
76 nn.SpatialBatchNormalization
1590.7774455175
77 nn.ReLU
1590.7774455175
78 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(320 -> 32, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(32 -> 64, 5x5, 2,2, 2,2)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
590.11505900137
79 nn.SpatialConvolution(320 -> 32, 1x1)
-2195.1067498252
80 nn.SpatialBatchNormalization
591.04037207016
81 nn.ReLU
591.04037207016
82 nn.SpatialConvolution(32 -> 64, 5x5, 2,2, 2,2)
-423.32057787105
83 nn.SpatialBatchNormalization
590.11505900137
84 nn.ReLU
590.11505900137
85 nn.Sequential {
[input -> (1) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
}
1937.2864550878
86 nn.SpatialMaxPooling(3x3, 2,2)
1937.2864550878
87 nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 192, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(640 -> 128, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(640 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
1419.2592620309
88 nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 192, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 32, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(640 -> 128, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (4): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(640 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
1419.2592620309
89 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(640 -> 96, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(96 -> 192, 3x3, 1,1, 1,1)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
450.65505624563
90 nn.SpatialConvolution(640 -> 96, 1x1)
-1807.8697174145
91 nn.SpatialBatchNormalization
494.42065476626
92 nn.ReLU
494.42065476626
93 nn.SpatialConvolution(96 -> 192, 3x3, 1,1, 1,1)
-6616.060830225
94 nn.SpatialBatchNormalization
450.65505624563
95 nn.ReLU
450.65505624563
96 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(640 -> 32, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
249.36065697856
97 nn.SpatialConvolution(640 -> 32, 1x1)
-971.76546653546
98 nn.SpatialBatchNormalization
116.89276184887
99 nn.ReLU
116.89276184887
100 nn.SpatialConvolution(32 -> 64, 5x5, 1,1, 2,2)
-1564.9647967899
101 nn.SpatialBatchNormalization
249.36065697856
102 nn.ReLU
249.36065697856
103 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Square
(2): nn.SpatialAveragePooling(3x3, 3,3)
(3): nn.MulConstant
(4): nn.Sqrt
}
(2): nn.SpatialConvolution(640 -> 128, 1x1)
(3): nn.SpatialBatchNormalization
(4): nn.ReLU
}
48.05579098314
104 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Square
(2): nn.SpatialAveragePooling(3x3, 3,3)
(3): nn.MulConstant
(4): nn.Sqrt
}
1936.6284440761
105 nn.Square
2488.1982231718
106 nn.SpatialAveragePooling(3x3, 3,3)
2488.1982171444
107 nn.MulConstant
2488.1982171444
108 nn.Sqrt
1936.6284440761
109 nn.SpatialConvolution(640 -> 128, 1x1)
-271.71464906633
110 nn.SpatialBatchNormalization
48.05579098314
111 nn.ReLU
48.05579098314
112 nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(640 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
671.18775782362
113 nn.SpatialConvolution(640 -> 256, 1x1)
-12056.546502997
114 nn.SpatialBatchNormalization
671.18775782362
115 nn.ReLU
671.18775782362
116 nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 160, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(160 -> 256, 3x3, 2,2, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 64, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(64 -> 128, 5x5, 2,2, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
}
... -> output
}
977.68043720257
117 nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 160, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(160 -> 256, 3x3, 2,2, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(640 -> 64, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(64 -> 128, 5x5, 2,2, 2,2)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
}
... -> output
}
977.68043720257
118 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(640 -> 160, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(160 -> 256, 3x3, 2,2, 1,1)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
223.0856897803
119 nn.SpatialConvolution(640 -> 160, 1x1)
-2699.4321808368
120 nn.SpatialBatchNormalization
659.31336273719
121 nn.ReLU
659.31336273719
122 nn.SpatialConvolution(160 -> 256, 3x3, 2,2, 1,1)
-2866.4377360148
123 nn.SpatialBatchNormalization
223.0856897803
124 nn.ReLU
223.0856897803
125 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(640 -> 64, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(64 -> 128, 5x5, 2,2, 2,2)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
188.83432870358
126 nn.SpatialConvolution(640 -> 64, 1x1)
-1479.4027583133
127 nn.SpatialBatchNormalization
252.81468722224
128 nn.ReLU
252.81468722224
129 nn.SpatialConvolution(64 -> 128, 5x5, 2,2, 2,2)
-576.61599449255
130 nn.SpatialBatchNormalization
188.83432870358
131 nn.ReLU
188.83432870358
132 nn.Sequential {
[input -> (1) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
}
565.76041871868
133 nn.SpatialMaxPooling(3x3, 2,2)
565.76041871868
134 nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(1024 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(1024 -> 96, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(1024 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
nan
135 nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(1024 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.Square
| (2): nn.SpatialAveragePooling(3x3, 3,3)
| (3): nn.MulConstant
| (4): nn.Sqrt
| }
| (2): nn.SpatialConvolution(1024 -> 96, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(1024 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
nan
136 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(1024 -> 96, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
274.44969951734
137 nn.SpatialConvolution(1024 -> 96, 1x1)
287.26738857804
138 nn.SpatialBatchNormalization
40.567358113825
139 nn.ReLU
40.567358113825
140 nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
-2398.1525074965
141 nn.SpatialBatchNormalization
274.44969951734
142 nn.ReLU
274.44969951734
143 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Square
(2): nn.SpatialAveragePooling(3x3, 3,3)
(3): nn.MulConstant
(4): nn.Sqrt
}
(2): nn.SpatialConvolution(1024 -> 96, 1x1)
(3): nn.SpatialBatchNormalization
(4): nn.ReLU
}
nan
144 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.Square
(2): nn.SpatialAveragePooling(3x3, 3,3)
(3): nn.MulConstant
(4): nn.Sqrt
}
521.42340321187
145 nn.Square
618.61728306977
146 nn.SpatialAveragePooling(3x3, 3,3)
618.61728176965
147 nn.MulConstant
618.61728176965
148 nn.Sqrt
521.42340321187
149 nn.SpatialConvolution(1024 -> 96, 1x1)
-5.9299795031548
150 nn.SpatialBatchNormalization
nan
151 nn.ReLU
nan
152 nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(1024 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
127.47477568407
153 nn.SpatialConvolution(1024 -> 256, 1x1)
-2194.8630077215
154 nn.SpatialBatchNormalization
127.47477568407
155 nn.ReLU
127.47477568407
156 nn.Inception @ nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(736 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.SpatialMaxPooling(3x3, 2,2)
| (2): nn.SpatialConvolution(736 -> 96, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(736 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
nan
157 nn.DepthConcat {
input
|`-> (1): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
| (1): nn.SpatialConvolution(736 -> 96, 1x1)
| (2): nn.SpatialBatchNormalization
| (3): nn.ReLU
| (4): nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
| (5): nn.SpatialBatchNormalization
| (6): nn.ReLU
| }
|`-> (2): nn.Sequential {
| [input -> (1) -> (2) -> (3) -> (4) -> output]
| (1): nn.SpatialMaxPooling(3x3, 2,2)
| (2): nn.SpatialConvolution(736 -> 96, 1x1)
| (3): nn.SpatialBatchNormalization
| (4): nn.ReLU
| }
|`-> (3): nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(736 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
... -> output
}
nan
158 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> (5) -> (6) -> output]
(1): nn.SpatialConvolution(736 -> 96, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
(4): nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
(5): nn.SpatialBatchNormalization
(6): nn.ReLU
}
nan
159 nn.SpatialConvolution(736 -> 96, 1x1)
nan
160 nn.SpatialBatchNormalization
nan
161 nn.ReLU
nan
162 nn.SpatialConvolution(96 -> 384, 3x3, 1,1, 1,1)
nan
163 nn.SpatialBatchNormalization
nan
164 nn.ReLU
nan
165 nn.Sequential {
[input -> (1) -> (2) -> (3) -> (4) -> output]
(1): nn.SpatialMaxPooling(3x3, 2,2)
(2): nn.SpatialConvolution(736 -> 96, 1x1)
(3): nn.SpatialBatchNormalization
(4): nn.ReLU
}
nan
166 nn.SpatialMaxPooling(3x3, 2,2)
112.96316097956
167 nn.SpatialConvolution(736 -> 96, 1x1)
126.45520535111
168 nn.SpatialBatchNormalization
nan
169 nn.ReLU
nan
170 nn.Sequential {
[input -> (1) -> (2) -> (3) -> output]
(1): nn.SpatialConvolution(736 -> 256, 1x1)
(2): nn.SpatialBatchNormalization
(3): nn.ReLU
}
nan
171 nn.SpatialConvolution(736 -> 256, 1x1)
nan
172 nn.SpatialBatchNormalization
nan
173 nn.ReLU
nan
174 nn.SpatialAveragePooling(3x3, 1,1)
nan
175 nn.View(736)
nan
176 nn.Linear(736 -> 128)
nan
177 nn.Normalize(2)
nan
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment