Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

R-package problem with "predict" function : array shapes mismatch #6919

Closed
ChrisBotella opened this issue Jul 3, 2017 · 4 comments
Closed

R-package problem with "predict" function : array shapes mismatch #6919

ChrisBotella opened this issue Jul 3, 2017 · 4 comments
Labels
R Windows

Comments

@ChrisBotella
Copy link

Environment info

Operating System: Windows 10
Package used : R
MXNet version: 0.10.1 (given by R). I installed the gpu version by following this tutorial https://gist.github.com/thirdwing/89aa9bfc588ade138496e6932072152c one month ago, and it was working fine until now.
R sessionInfo():
R version 3.3.2 (2016-10-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows >= 8 x64 (build 9200)

Error Message:

[18:18:32] d:\program files (x86)\jenkins\workspace\mxnet\mxnet\dmlc-core\include\dmlc./logging.h:304: [18:18:32] d:\program files (x86)\jenkins\workspace\mxnet\mxnet\src\operator\tensor../elemwise_op_common.h:33: Check failed: assign(&dattr, (*vec)[i]) Incompatible attr in node _mul11 at 1-th input: expected (89,), got (89,10)
Error in symbol$infer.shape(list(...)) :
Error in operator _mul11: [18:18:32] d:\program files (x86)\jenkins\workspace\mxnet\mxnet\src\operator\tensor../elemwise_op_common.h:33: Check failed: assign(&dattr, (*vec)[i]) Incompatible attr in node _mul11 at 1-th input: expected (89,), got (89,10)

Minimum reproducible example

# This example is a multi-label regression on a generalized linear model with poisson likelihood
# we simulate random data
p = 32
n_label = 10
n = 89
X = matrix(runif(p*n,0,1),p,n)
y = matrix(round(rpois(n_label*n,10)),n_label,n)
# model building
data <- mx.symbol.Variable("data")
label=mx.symbol.Variable('label')
# we create a parametrized linear combination of input variables in fc
fc <- mx.symbol.FullyConnected(data, num_hidden=n_label,name="fc")
# As the loss, we write the Negative Log-Likelihood associated with an exponential link function
vecto.symb = mx.symbol.MakeLoss(data= mx.symbol.exp(fc) - label * fc , name="poisson")
devices = mx.gpu(0)
model = mx.model.FeedForward.create(symbol = vecto.symb,ctx = devices, X=X,y=y,num.round=5,array.layout="colmajor",learning.rate=0.01, optimizer="sgd",initializer=mx.init.normal(0.03),array.batch.size=20)
p = predict(model,X,array.layout="colmajor")

What have you tried to solve it?

The error appears when using the predict function. The same code works fine on another machine (same Windows 10, same R version) which has the classic cpu R install (used install.packages) of mxnet.

So the error might be an installation problem with the gpu version? or it might be related to #113 ?
@thirdwing
Copy link
Contributor

It seems the Windows pkg needs update.

On my Linux machine:

> library(mxnet)
> p = 32
> n_label = 10
> n = 89
> X = matrix(runif(p*n,0,1),p,n)
> y = matrix(round(rpois(n_label*n,10)),n_label,n)
> data <- mx.symbol.Variable("data")
> label=mx.symbol.Variable('label')
> fc <- mx.symbol.FullyConnected(data, num_hidden=n_label,name="fc")
> vecto.symb = mx.symbol.MakeLoss(data= mx.symbol.exp(fc) - label * fc , name="poisson")
> devices = mx.gpu(0)
> model = mx.model.FeedForward.create(symbol = vecto.symb,ctx = devices, X=X,y=y,num.round=5,array.layout="colmajor",learning.rate=0.01, optimizer="sgd",initializer=mx.init.normal(0.03),array.batch.size=20)
Start training with 1 devices
> p = predict(model,X,array.layout="colmajor")
> p
          [,1]      [,2]      [,3]     [,4]     [,5]     [,6]      [,7]     [,8]     [,9]    [,10]
 [1,] 7.816670  8.157823 10.474072 11.39288 7.779306 6.970769  9.895404 12.01964 7.290720 8.944089
 [2,] 7.453233  8.445386 10.373491 12.59584 7.349095 6.664856 10.164314 11.51193 6.594102 8.864347
 [3,] 8.928808 10.708864 10.548546 13.22053 7.846766 7.146811 10.764274 13.85162 7.298235 9.293221
 [4,] 8.230606  9.123090 11.217604 13.23091 8.138767 7.019768 10.248020 14.35250 7.356753 8.135243
 [5,] 8.071321  8.288631 10.508270 12.12168 7.196826 6.733716  9.084327 12.06030 6.710617 8.396719
 [6,] 8.482371 10.366955  9.545642 11.35876 7.762624 6.653240 11.191854 13.40291 7.470777 8.287533
 [7,] 7.665237  9.516909 10.175713 12.69094 7.598200 6.970988 11.616826 13.16461 6.503124 9.161448
 [8,] 7.701565  9.458061  9.218641 12.51689 7.938196 6.988401 11.182857 12.04601 7.121068 8.645683
 [9,] 7.566068  8.341774  8.861707 10.38127 6.559480 6.436257  9.217819 10.99246 6.040694 7.019975
[10,] 8.520127  9.488525 10.543437 12.88380 7.868061 7.011839 10.414857 12.94717 7.134251 9.127591
          [,11]     [,12]    [,13]    [,14]    [,15]     [,16]     [,17]     [,18]     [,19]    [,20]
 [1,]  9.168724  8.523341 7.678204 7.943961 7.256366 10.397227  9.454224  9.628304 10.630404 9.333558
 [2,]  8.994869 10.060082 8.047057 7.978253 8.053446  9.684225  8.919069  9.212090 10.225703 9.098025
 [3,]  9.813600 10.093615 8.200279 7.718523 7.769364 11.378102  9.855852 10.003072 10.997955 9.812004
 [4,]  9.079152 11.014524 7.525928 8.309593 6.809573 12.059962 10.649134  9.013678 12.408504 9.326276
 [5,]  7.977436  9.912253 6.456532 7.237380 6.385749 10.745084  8.633645  8.501617 10.114634 7.791702
 [6,]  9.058272  8.981851 7.984339 7.783956 7.703894 10.208138 10.439919 11.094374 11.401168 8.967443
 [7,]  9.031200 10.548182 9.062616 8.302197 7.647370 10.454777 10.687029  9.221517 13.639665 8.928701
 [8,] 10.083397  9.335425 8.125834 9.245922 7.748134 10.652124  9.383847  9.822195 10.822229 8.678563
 [9,]  8.763423  8.120615 6.653493 7.085348 6.808999  8.905265  8.267180  8.320374  9.960668 8.852212
[10,]  9.612265  9.996041 8.390809 8.186895 7.708246 10.205514  9.550326  9.781266 12.611280 9.213140
          [,21]     [,22]     [,23]    [,24]     [,25]    [,26]     [,27]    [,28]     [,29]    [,30]
 [1,] 11.308876  9.288301  9.569324 12.46846 10.172219 8.790600  9.483993 13.28197 10.465868 11.33658
 [2,] 11.253362  8.632038  9.748217 11.87590 10.855169 9.407741 10.046068 11.42039  9.945236 11.29840
 [3,] 12.477365  9.152213 11.524874 15.02225 11.458111 9.633025  9.971084 14.89848 10.365815 12.15181
 [4,] 11.579143 10.519217 11.190501 14.75211 12.171769 9.657764 11.083182 14.42005 11.121315 11.32189
 [5,] 10.637225  9.505926  9.946135 12.60800 11.126405 8.921020  9.440014 13.36149  9.041626 10.35690
 [6,] 12.291706  8.437302 10.696656 14.43274 10.379085 9.411674  9.331091 12.90256  9.731915 11.89710
 [7,] 12.905643  9.070361 11.840013 14.30707 11.611399 9.670755  9.901644 14.70455  9.644423 11.26303
 [8,] 11.293889  8.842366 10.220294 14.22478 10.802110 8.745304 10.874222 14.66447  9.923064 11.58790
 [9,]  9.354656  7.657266  8.565167 11.63647  9.593769 8.502037  9.264493 11.46476  9.335666 10.89800
[10,] 11.988859  9.438472 10.797145 14.80075 11.518509 9.250444  9.859594 13.98825 10.614679 11.72729
          [,31]    [,32]    [,33]     [,34]     [,35]    [,36]    [,37]    [,38]     [,39]    [,40]
 [1,]  9.866580 10.49420 8.642398  8.790415  9.733511 11.99423 6.756829 7.628892  9.427886 7.902532
 [2,]  8.441378 10.91738 7.268434  9.354983  8.620921 11.54677 6.560183 7.412773  9.592722 7.433565
 [3,] 11.526748 12.11366 9.255115 10.099893 11.629895 13.73594 7.003594 8.172505 11.263753 7.457634
 [4,]  9.954642 11.29973 9.896523 10.200948  9.685892 14.23606 7.303020 8.677130 10.437410 7.933517
 [5,]  9.637197 10.60180 8.510577  9.413569  8.995958 12.14189 6.585538 8.386744  9.185835 7.152886
 [6,] 11.354940 10.81478 9.096291  9.356621 10.409036 13.30373 6.882611 7.748830 10.529068 6.938926
 [7,] 10.151248 11.24501 9.504297 10.656692  9.294201 14.29821 8.126706 7.991077 10.253646 7.316355
 [8,]  9.457038 12.77455 9.239510  9.746669 11.279565 14.08356 6.649528 9.087983  9.574911 8.240922
 [9,]  8.786914  9.41922 7.200623  8.183215  8.312158 11.07285 5.736485 6.726950  8.709372 6.553879
[10,]  9.437313 11.99361 9.123729  9.999371  9.689675 12.84750 7.461349 8.532560 10.222152 7.424592
          [,41]    [,42]     [,43]    [,44]     [,45]     [,46]     [,47]    [,48]     [,49]    [,50]
 [1,]  8.649339 5.314861 12.621219 5.514723 10.003050  9.503062 11.746906 11.38585 11.742462 8.583298
 [2,]  7.704547 5.672434 10.236336 5.900691  9.757183  8.363594 11.409889 10.86170 10.496987 7.399814
 [3,] 10.187313 5.949660 12.824229 6.548615 11.043896 11.143600 12.783177 13.57207 12.047352 8.578787
 [4,]  8.473114 5.517111 12.646370 6.752975 11.089304 10.094254 12.687654 14.06325 12.750614 8.981051
 [5,]  8.205828 5.278526 11.942309 6.155215 10.293086  9.945828 11.597883 11.71625 10.327518 7.684636
 [6,]  9.521443 6.392505 12.312798 6.588591 11.302396  9.797333 11.525074 12.75276 12.482216 8.209350
 [7,]  8.503182 6.182213 13.231131 6.001590 11.974772 10.321801 14.011294 11.64822 13.432488 7.991267
 [8,]  9.458187 6.004597 12.362864 5.971149 11.299578  8.827434 12.157065 12.39539 11.267076 7.882904
 [9,]  7.755996 5.199047  9.454911 5.931271  9.107282  7.602388  9.795255 11.51299  9.897947 7.584461
[10,]  9.022410 6.131191 11.992719 6.484079 11.273059  9.595859 12.916764 12.10417 11.582853 8.536198
         [,51]     [,52]    [,53]    [,54]    [,55]     [,56]    [,57]    [,58]     [,59]    [,60]
 [1,] 9.244983  9.026376 11.60768 7.445289 12.20319  9.007307 8.202882 11.69678 11.291792 11.53786
 [2,] 8.126652  8.328223 13.41882 7.544335 10.47258  9.024238 9.460758 11.10555 10.212744 10.30471
 [3,] 8.480664 10.294396 13.88592 8.045505 12.07821  9.582493 9.304089 11.37732 12.087280 13.55252
 [4,] 9.313348  9.124362 13.88364 9.203846 12.22964  9.436004 8.649411 12.50475 11.735001 12.62620
 [5,] 7.924311  8.885804 12.02092 7.150859 10.57416  8.598110 8.200793 10.52620 10.053449 10.77155
 [6,] 8.096569  9.666099 12.58955 7.703288 12.29899  8.817454 8.296067 10.80231 10.658768 11.92008
 [7,] 8.842400  9.423360 13.69213 9.047062 11.73560  9.024308 8.826710 11.14274 12.560871 12.56917
 [8,] 9.414310  9.070774 12.86080 7.420550 12.20953 10.125348 9.079483 12.66835 11.415252 12.35455
 [9,] 7.392401  7.960190 11.09324 6.392342 10.30098  8.214180 8.085503 10.24684  9.695411 10.35095
[10,] 9.260792  9.340340 14.07300 8.074622 12.85251  9.981878 9.257211 11.91849 11.715885 12.43994
          [,61]     [,62]     [,63]     [,64]     [,65]     [,66]     [,67]     [,68]    [,69]    [,70]
 [1,]  8.964652 10.141632  9.096560  8.861149 10.476036  9.345021 10.259480  9.010168 11.57214 6.659705
 [2,]  8.534002 10.257945  7.253962  8.705607  9.481013  7.980782 10.571997  8.232688 10.29385 6.539469
 [3,] 10.711432 10.380169  8.970151  9.116873 11.060877 11.656600 11.390860 11.598317 12.59681 7.431988
 [4,] 10.769454 10.100412 10.068944 10.034698 11.652020 10.033549 11.318120 10.702707 12.96376 7.485595
 [5,]  9.410174  9.412863  9.099507  9.225079 10.146431 10.289325 11.106381 10.103276 11.16685 6.908048
 [6,]  9.705215  9.786076  8.753164  8.416007 11.158924  9.879974 11.187751 10.781698 11.69304 7.059194
 [7,]  9.968441 10.295626  9.260633  9.638412 10.761553  9.065764 11.900574  9.745747 11.93590 6.707314
 [8,]  9.288058  9.857211  8.992968  9.769420 10.898926  9.228750 10.567677  8.946614 11.19650 7.743548
 [9,]  8.392508  8.878320  7.380243  8.299747  8.773889  8.152227  8.889181  8.789065 10.46103 6.591502
[10,] 10.565228 10.766788  9.772991  9.537443 11.391335  9.922419 11.529438 10.707608 12.25396 7.643187
         [,71]    [,72]     [,73]    [,74]     [,75]    [,76]     [,77]    [,78]    [,79]     [,80]
 [1,] 8.496582 5.932070  8.736745 7.127186 10.683019 11.51356  8.878646 7.619807 11.38468 11.048352
 [2,] 7.593750 6.164192  8.259268 6.310826  9.306476 11.35056  9.147801 7.353465 10.92469  9.857203
 [3,] 9.020813 6.697681 10.578875 8.397291 11.469223 13.95665  9.171689 8.823219 13.22135 12.855165
 [4,] 8.484274 6.673910  9.801679 6.904182 11.306722 12.91221  9.277265 9.201324 12.87133 11.594430
 [5,] 8.706759 6.278548  8.878038 6.869167 10.205997 11.35329  8.014498 8.201095 11.34289 10.734923
 [6,] 9.373224 6.283204 10.551237 7.767508 10.409093 13.09805  9.084136 7.589930 11.67227 11.563037
 [7,] 8.381074 6.922416 10.277338 7.866474 11.223840 13.72707 10.068846 8.340858 10.85645 10.862444
 [8,] 8.100209 6.363280  9.032307 7.740890 11.322384 11.65092  8.911162 7.813134 12.39743 11.608423
 [9,] 6.830784 5.390416  7.995050 6.114217  9.249428 10.38231  7.801753 7.293544 10.63287  9.385077
[10,] 8.761749 6.808009  9.693430 7.363374 10.770193 12.13806  9.655453 8.921905 11.41752 11.697877
         [,81]    [,82]    [,83]    [,84]     [,85]     [,86]    [,87]    [,88]     [,89]
 [1,] 8.319080 9.214132 7.316243 8.342993  8.825691 10.382159 6.941805 8.265296 11.088474
 [2,] 7.492801 7.521801 7.128818 8.230494  7.547138 11.223130 6.734413 8.538315  9.608241
 [3,] 8.821959 9.503194 7.671940 9.203848 10.735788 13.862556 7.180660 8.137114  9.824462
 [4,] 8.719627 9.177809 8.077137 9.009045 10.390970 12.364767 7.818877 8.203750 10.473236
 [5,] 7.458200 8.372662 7.605329 8.418344  9.782778 11.312627 6.777676 7.380335  9.437925
 [6,] 9.505312 9.496328 7.079411 8.963075 11.295750 12.375880 7.207112 7.754936  9.533134
 [7,] 9.581206 9.374516 7.761055 9.984800  9.554749 12.038147 6.906099 8.259149 10.723992
 [8,] 8.620052 9.202593 7.301843 9.875078  9.660048 12.118854 7.423788 7.825188 11.017103
 [9,] 7.345890 7.959776 6.273556 7.370767  8.206044  9.588604 6.565289 7.717051  8.245582
[10,] 8.573742 8.522071 7.391146 9.527450  9.333231 11.119854 7.414830 8.066483 10.641735
>

@ChrisBotella
Copy link
Author

@thirdwing , So you reckon it should work if I re-do the installation (https://gist.github.com/thirdwing/89aa9bfc588ade138496e6932072152c) with the last update of windows build (by downloading the last version of mxnet_x64_vc14_gpu.7z and prebuildbase_win10_x64_vc14.7z ) ?

@thirdwing
Copy link
Contributor

It should work if you use the latest code.

You can also wait for a while. I will update the Windows pkg.

@thirdwing
Copy link
Contributor

The prebuilt pkgs have been updated.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees
No one assigned
Labels
R Windows
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@thirdwing @ChrisBotella