An error in BinaryNet_SVHN_Model.lua

[kwakai3]

I attempted to execute the BinaryNet_SVHN.lua but got an error "bad argument #4 to '?' (cannot convert 'int *' to 'int')" as follows. Re-install of nn, cutorch, cunn and cudnn did not work. Do you have any suggestions?
Thank you in advance for your cooperation.

BinaryNet user$ th Main_BinaryNet_SVHN.lua -network BinaryNet_SVHN_Model
sh: mk1ir: command not found
0
6406494
6397632
6406494
/Users/user/BinaryNet/PreProcData/SVHN/rgbGCN_LCN_valid.t7
applying GCN preprocessing
[================== 598388/598388 ============>]ETA: 0ms | Step: 0ms
applying LeCunLCN preprocessing
[================== 598388/598388 ============>]ETA: 0ms | Step: 0ms
applying GCN preprocessing
[================== 6000/6000 ================>]ETA: 0ms | Step: 0ms
applying LeCunLCN preprocessing
[================== 6000/6000 ================>]ETA: 0ms | Step: 0ms
applying GCN preprocessing
[================== 26032/26032 ==============>]ETA: 0ms | Step: 0ms
applying LeCunLCN preprocessing
[================== 26032/26032 ==============>]ETA: 0ms | Step: 0ms
[program started on Fri Jun 17 05:39:50 2016]
[command line arguments]
stcWeights false
LR 0.0078125
modelsFolder ./Models/
batchSize 200
optimization adam
preProcDir /Users/user/BinaryNet/PreProcData/SVHN
network ./Models/BinaryNet_SVHN_Model
stcNeurons true
constBatchSize false
LRDecay 0
whiten false
augment false
load
nGPU 1
dp_prepro true
format rgb
save /Users/user/BinaryNet/Results/FriJun1705:33:592016
dataset SVHN
normalization simple
devid 1
visualize 1
type cuda
threads 8
SBN true
momentum 0
weightDecay 0
runningVal true
epoch -1
[----------------------]
==> Network
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) -> (25) -> (26) -> (27) -> (28) -> (29) -> (30) -> (31) -> (32) -> (33) -> (34) -> (35) -> (36) -> (37) -> (38) -> output: cudnnBinarySpatialConvolution(3 -> 64, 3x3, 1,1, 1,1)
(2): SpatialBatchNormalizationShiftPow2
(3): nn.HardTanh
(4): BinarizedNeurons
(5): cudnnBinarySpatialConvolution(64 -> 64, 3x3, 1,1, 1,1)
(6): cudnn.SpatialMaxPooling(2x2, 2,2)
(7): SpatialBatchNormalizationShiftPow2
(8): nn.HardTanh
(9): BinarizedNeurons
(10): cudnnBinarySpatialConvolution(64 -> 128, 3x3, 1,1, 1,1)
(11): SpatialBatchNormalizationShiftPow2
(12): nn.HardTanh
(13): BinarizedNeurons
(14): cudnnBinarySpatialConvolution(128 -> 128, 3x3, 1,1, 1,1)
(15): cudnn.SpatialMaxPooling(2x2, 2,2)
(16): SpatialBatchNormalizationShiftPow2
(17): nn.HardTanh
(18): BinarizedNeurons
(19): cudnnBinarySpatialConvolution(128 -> 256, 3x3, 1,1, 1,1)
(20): SpatialBatchNormalizationShiftPow2
(21): nn.HardTanh
(22): BinarizedNeurons
(23): cudnnBinarySpatialConvolution(256 -> 256, 3x3, 1,1, 1,1)
(24): cudnn.SpatialMaxPooling(2x2, 2,2)
(25): SpatialBatchNormalizationShiftPow2
(26): nn.HardTanh
(27): BinarizedNeurons
(28): nn.View(4096)
(29): BinaryLinear(4096 -> 1024)
(30): BatchNormalizationShiftPow2
(31): nn.HardTanh
(32): BinarizedNeurons
(33): BinaryLinear(1024 -> 1024)
(34): BatchNormalizationShiftPow2
(35): nn.HardTanh
(36): BinarizedNeurons
(37): BinaryLinear(1024 -> 10)
(38): nn.BatchNormalization
}
==>6406494 Parameters
==> Loss
SqrtHingeEmbeddingCriterion

==> Starting Training

Epoch 1
/Users/user/torch/install/bin/luajit: /Users/user/torch/install/share/lua/5.1/nn/Container.lua:67:
In 1 module of nn.Sequential:
/Users/user/torch/install/share/lua/5.1/cudnn/init.lua:55: bad argument #4 to '?' (cannot convert 'int *' to 'int')
stack traceback:
[builtin#173]: at 0x0107d55450
/Users/user/torch/install/share/lua/5.1/cudnn/init.lua:55: in function 'errcheck'
.../user/BinaryNet/Models/cudnnBinarySpatialConvolution.lua:71: in function 'resetWeightDescriptors'
...torch/install/share/lua/5.1/cudnn/SpatialConvolution.lua:364: in function 'updateOutput'
.../user/BinaryNet/Models/cudnnBinarySpatialConvolution.lua:146: in function <.../user/BinaryNet/Models/cudnnBinarySpatialConvolution.lua:142>
[C]: in function 'xpcall'
/Users/user/torch/install/share/lua/5.1/nn/Container.lua:63: in function 'rethrowErrors'
/Users/user/torch/install/share/lua/5.1/nn/Sequential.lua:44: in function 'forward'
Main_BinaryNet_SVHN.lua:215: in function 'Train'
Main_BinaryNet_SVHN.lua:279: in main chunk
[C]: in function 'dofile'
...user/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:131: in main chunk
[C]: at 0x0107cd61a0

WARNING: If you see a stack trace below, it doesn't point to the place where this error occured. Please use only the one above.
stack traceback:
[C]: in function 'error'
/Users/user/torch/install/share/lua/5.1/nn/Container.lua:67: in function 'rethrowErrors'
/Users/user/torch/install/share/lua/5.1/nn/Sequential.lua:44: in function 'forward'
Main_BinaryNet_SVHN.lua:215: in function 'Train'
Main_BinaryNet_SVHN.lua:279: in main chunk
[C]: in function 'dofile'
...user/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:131: in main chunk
[C]: at 0x0107cd61a0

[kwakai3]

I could fix the error by modification of cudnnBinarySpatialConvolution.lua as follows.

Lines 71 to 73:
Original
errcheck('cudnnSetFilterNdDescriptor', self.weightDesc[0],
'CUDNN_DATA_FLOAT', 4,
desc:data());

Modified
errcheck('cudnnSetFilterNdDescriptor', self.weightDesc[0],
'CUDNN_DATA_FLOAT', 'CUDNN_TENSOR_NCHW', 4,
desc:data());

[MinLiAmoy]

same error occurs when using "nn" library on CPU, but BinarySpatialConvolution.lua doesn't have such errcheck code. How to modify it?
the error is shown below:
Epoch 1
/home/ml/torch/install/bin/luajit: /home/ml/torch/install/share/lua/5.1/nn/Container.lua:67:
In 23 module of nn.Sequential:
/home/ml/torch/install/share/lua/5.1/nn/THNN.lua:110: bad argument #8 to 'v' (cannot convert 'struct THFloatTensor *' to 'int')
stack traceback:
[C]: in function 'v'
/home/ml/torch/install/share/lua/5.1/nn/THNN.lua:110: in function 'SpatialConvolutionMM_updateGradInput'
...CLA/work/BinaryNet-1/Models/BinarySpatialConvolution.lua:150: in function 'updateGradInput'
/home/ml/torch/install/share/lua/5.1/nn/Module.lua:31: in function </home/ml/torch/install/share/lua/5.1/nn/Module.lua:29>
[C]: in function 'xpcall'
/home/ml/torch/install/share/lua/5.1/nn/Container.lua:63: in function 'rethrowErrors'
/home/ml/torch/install/share/lua/5.1/nn/Sequential.lua:84: in function 'backward'
Main_BinaryNet_Cifar10.lua:227: in function 'opfunc'
./adaMax_binary_clip_shift.lua:38: in function 'adaMax_binary_clip_shift'
Main_BinaryNet_Cifar10.lua:231: in function 'Train'
Main_BinaryNet_Cifar10.lua:270: in main chunk
[C]: in function 'dofile'
...e/ml/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
[C]: at 0x00405d50

WARNING: If you see a stack trace below, it doesn't point to the place where this error occurred. Please use only the one above.
stack traceback:
[C]: in function 'error'
/home/ml/torch/install/share/lua/5.1/nn/Container.lua:67: in function 'rethrowErrors'
/home/ml/torch/install/share/lua/5.1/nn/Sequential.lua:84: in function 'backward'
Main_BinaryNet_Cifar10.lua:227: in function 'opfunc'
./adaMax_binary_clip_shift.lua:38: in function 'adaMax_binary_clip_shift'
Main_BinaryNet_Cifar10.lua:231: in function 'Train'
Main_BinaryNet_Cifar10.lua:270: in main chunk
[C]: in function 'dofile'
...e/ml/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
[C]: at 0x00405d50