Skip to content

Instantly share code, notes, and snippets.

@WolframRhodium
Last active October 31, 2018 09:47
Show Gist options
  • Save WolframRhodium/1868cac2fb0354950ff6e2a9e40ab65e to your computer and use it in GitHub Desktop.
Save WolframRhodium/1868cac2fb0354950ff6e2a9e40ab65e to your computer and use it in GitHub Desktop.
AOD_net (fix preprocessing)
{
"nodes": [
{
"op": "null",
"name": "data",
"inputs": []
},
{
"op": "_plus_scalar",
"name": "_plusscalar0",
"attrs": {"scalar": "-0.5"},
"inputs": [[0, 0, 0]]
},
{
"op": "_mul_scalar",
"name": "_mulscalar0",
"attrs": {"scalar": "2"},
"inputs": [[1, 0, 0]]
},
{
"op": "null",
"name": "learned_0",
"attrs": {"__shape__": "(3, 3, 1, 1)"},
"inputs": []
},
{
"op": "null",
"name": "learned_1",
"attrs": {"__shape__": "(3,)"},
"inputs": []
},
{
"op": "Convolution",
"name": "convolution0",
"attrs": {
"dilate": "(1, 1)",
"kernel": "(1, 1)",
"layout": "NCHW",
"no_bias": "False",
"num_filter": "3",
"num_group": "1",
"pad": "(0, 0)",
"stride": "(1, 1)"
},
"inputs": [[2, 0, 0], [3, 0, 0], [4, 0, 0]]
},
{
"op": "relu",
"name": "relu0",
"inputs": [[5, 0, 0]]
},
{
"op": "null",
"name": "learned_2",
"attrs": {"__shape__": "(3, 3, 3, 3)"},
"inputs": []
},
{
"op": "null",
"name": "learned_3",
"attrs": {"__shape__": "(3,)"},
"inputs": []
},
{
"op": "Convolution",
"name": "convolution1",
"attrs": {
"dilate": "(1, 1)",
"kernel": "(3, 3)",
"layout": "NCHW",
"no_bias": "False",
"num_filter": "3",
"num_group": "1",
"pad": "(1, 1)",
"stride": "(1, 1)"
},
"inputs": [[6, 0, 0], [7, 0, 0], [8, 0, 0]]
},
{
"op": "relu",
"name": "relu1",
"inputs": [[9, 0, 0]]
},
{
"op": "Concat",
"name": "concat0",
"attrs": {
"dim": "1",
"num_args": "2"
},
"inputs": [[6, 0, 0], [10, 0, 0]]
},
{
"op": "null",
"name": "learned_4",
"attrs": {"__shape__": "(3, 6, 5, 5)"},
"inputs": []
},
{
"op": "null",
"name": "learned_5",
"attrs": {"__shape__": "(3,)"},
"inputs": []
},
{
"op": "Convolution",
"name": "convolution2",
"attrs": {
"dilate": "(1, 1)",
"kernel": "(5, 5)",
"layout": "NCHW",
"no_bias": "False",
"num_filter": "3",
"num_group": "1",
"pad": "(2, 2)",
"stride": "(1, 1)"
},
"inputs": [[11, 0, 0], [12, 0, 0], [13, 0, 0]]
},
{
"op": "relu",
"name": "relu2",
"inputs": [[14, 0, 0]]
},
{
"op": "Concat",
"name": "concat1",
"attrs": {
"dim": "1",
"num_args": "2"
},
"inputs": [[10, 0, 0], [15, 0, 0]]
},
{
"op": "null",
"name": "learned_6",
"attrs": {"__shape__": "(3, 6, 7, 7)"},
"inputs": []
},
{
"op": "null",
"name": "learned_7",
"attrs": {"__shape__": "(3,)"},
"inputs": []
},
{
"op": "Convolution",
"name": "convolution3",
"attrs": {
"dilate": "(1, 1)",
"kernel": "(7, 7)",
"layout": "NCHW",
"no_bias": "False",
"num_filter": "3",
"num_group": "1",
"pad": "(3, 3)",
"stride": "(1, 1)"
},
"inputs": [[16, 0, 0], [17, 0, 0], [18, 0, 0]]
},
{
"op": "relu",
"name": "relu3",
"inputs": [[19, 0, 0]]
},
{
"op": "Concat",
"name": "concat2",
"attrs": {
"dim": "1",
"num_args": "4"
},
"inputs": [[6, 0, 0], [10, 0, 0], [15, 0, 0], [20, 0, 0]]
},
{
"op": "null",
"name": "learned_8",
"attrs": {"__shape__": "(3, 12, 3, 3)"},
"inputs": []
},
{
"op": "null",
"name": "learned_9",
"attrs": {"__shape__": "(3,)"},
"inputs": []
},
{
"op": "Convolution",
"name": "convolution4",
"attrs": {
"dilate": "(1, 1)",
"kernel": "(3, 3)",
"layout": "NCHW",
"no_bias": "False",
"num_filter": "3",
"num_group": "1",
"pad": "(1, 1)",
"stride": "(1, 1)"
},
"inputs": [[21, 0, 0], [22, 0, 0], [23, 0, 0]]
},
{
"op": "relu",
"name": "relu4",
"inputs": [[24, 0, 0]]
},
{
"op": "broadcast_mul",
"name": "broadcast_mul0",
"inputs": [[25, 0, 0], [2, 0, 0]]
},
{
"op": "broadcast_sub",
"name": "broadcast_sub0",
"inputs": [[26, 0, 0], [25, 0, 0]]
},
{
"op": "_plus_scalar",
"name": "_plusscalar0",
"attrs": {"scalar": "1.0"},
"inputs": [[27, 0, 0]]
},
{
"op": "relu",
"name": "relu5",
"inputs": [[28, 0, 0]]
}
],
"arg_nodes": [
0,
3,
4,
7,
8,
12,
13,
17,
18,
22,
23
],
"node_row_ptr": [
0,
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
],
"heads": [[29, 0, 0]],
"attrs": {"mxnet_version": ["int", 10300]}
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment