require 'cunn' has some problems. (CUDA8)

[Tao]

when I utilize torch and call the command  require 'cunn', I have the following problems, it seems strange. Does anyone have a similar situation and have a good solution？ Thank you。

aHalfMarginCriterion_updateGradInput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfMultiMarginCriterion_updateOutput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfMultiMarginCriterion_updateGradInput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfPReLU_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSpatialConvolutionLocal_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSpatialConvolutionMM_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSpatialCrossMapLRN_updateOutput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSpatialCrossMapLRN_updateGradInput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSpatialDilatedConvolution_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSpatialFullConvolution_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSoftPlus_updateOutput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSoftPlus_updateGradInput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSoftShrink_updateOutput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSoftShrink_updateGradInput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfSqrt_updateOutput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfTemporalConvolution_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfThreshold_updateOutput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfThreshold_updateGradInput/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfVolumetricConvolution_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfVolumetricDilatedConvolution_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

not found: THNN_CudaHalfVolumetricFullConvolution_accGradParameters/home/usr/torch/install/share/lua/5.2/nn/THNN.lua:108: NYI: call arg type

I already update the cutorch, cunn, nn. It does not make sense. 

[Tao]

Does someone know the reason why I have such a problem?

在 2016年12月7日星期三 UTC+2下午11:21:13，Tao写道：

[Tao]

I found a solution. However, I still did not know why it can work. I remove torch and install it again. Then this problem is never shown, 

在 2016年12月7日星期三 UTC+2下午11:21:13，Tao写道：

when I utilize torch and call the command  require 'cunn', I have the following problems, it seems strange. Does anyone have a similar situation and have a good solution？ Thank you。

[Mina Khoshdeli]

Reinstalling torch solves this for me. I still would like to know what caused the problem!
Thnak You!

[Tao]

I still have no idea why this will happen. However, I never have this problem when I use other versions of CUDA. So I guess it may be caused by the CUDA 8. If you find the final solution, please let us know. Thank you. 

在 2016年12月7日星期三 UTC+2下午11:21:13，Tao写道：

when I utilize torch and call the command  require 'cunn', I have the following problems, it seems strange. Does anyone have a similar situation and have a good solution？ Thank you。