torch运行错误libcudnn && Unsupported HDF5 version: 1.10.2错误

（一）torch运行程序遇到错误libcudnn

Please install CuDNN from https://developer.nvidia.com/cuDNN
Then make sure files named as libcudnn.so.7 or libcudnn.7.dylib are placed in
your library load path (for example /usr/local/lib , or manually add a path to LD_LIBRARY_PATH)

Alternatively, set the path to libcudnn.so.7 or libcudnn.7.dylib
to the environment variable CUDNN_PATH and rerun torch.
For example: export CUDNN_PATH = "/usr/local/cuda/lib64/libcudnn.so.7"

stack traceback:
	[C]: in function 'error'
	...eugraphubuntu/torch/install/share/lua/5.1/trepl/init.lua:389: in function 'require'
	main_test_bound_realsense.lua:4: in main chunk
	[C]: in function 'dofile'
	...untu/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
	[C]: at 0x00405d50
/home/seugraphubuntu/torch/install/bin/luajit: ...eugraphubuntu/torch/install/share/lua/5.1/trepl/init.lua:389: ...eugraphubuntu/torch/install/share/lua/5.1/trepl/init.lua:389: ...seugraphubuntu/torch/install/share/lua/5.1/cudnn/ffi.lua:1753: 'libcudnn (R7\) not found in library path.
Please install CuDNN from https://developer.nvidia.com/cuDNN
Then make sure files named as libcudnn.so.7 or libcudnn.7.dylib are placed in
your library load path (for example /usr/local/lib , or manually add a path to LD_LIBRARY_PATH)

Alternatively, set the path to libcudnn.so.7 or libcudnn.7.dylib
to the environment variable CUDNN_PATH and rerun torch.
For example: export CUDNN_PATH = "/usr/local/cuda/lib64/libcudnn.so.7"

stack traceback:
	[C]: in function 'error'
	...eugraphubuntu/torch/install/share/lua/5.1/trepl/init.lua:389: in function 'require'
	main_test_realsense.lua:4: in main chunk
	[C]: in function 'dofile'
	...untu/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
	[C]: at 0x00405d50

按照提示解决方案去做，并没有解决问题

我是cuda8.0+cudnn6.0+torch7.0

正确解决方案

运行以下代码

git clone https://github.com/soumith/cudnn.torch.git -b R6 && cd cudnn.torch && luarocks make cudnn-scm-1.rockspec

如果是cudnn7.0的，把R6改成R7即可。

（二）Unsupported HDF5 version: 1.10.2错误

错误代码大致如下

torch/install/share/lua/5.1/trepl/init.lua:389: /home/seugraphubuntu/.luarocks/share/lua/5.1/hdf5/ffi.lua:71: Unsupported HDF5 version: 1.10.2
stack traceback:
	[C]: in function 'error'
	...eugraphubuntu/torch/install/share/lua/5.1/trepl/init.lua:389: in function 'require'
	main_test_realsense.lua:13: in main chunk
	[C]: in function 'dofile'
	...untu/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
	[C]: at 0x00405d50

解决方案：

Edit: /Users//torch/install/share/lua/5.2/hdf5/ffi.lua
change
" maj[0] ~= 1 or min[0] ~= 8"
to
" maj[0] ~= 1 or min[0] ~= 10"

解决后还有问题。。。。一并记下。。。

（三）错误提示

.luarocks/share/lua/5.1/hdf5/file.lua:10: HDF5File.__init() requires a fileID - perhaps you want HDF5File.create()?
stack traceback:
	[C]: in function 'assert'
	/home/seugraphubuntu/.luarocks/share/lua/5.1/hdf5/file.lua:10: in function '__init'
	/home/seugraphubuntu/.luarocks/share/lua/5.1/torch/init.lua:91: in function </home/seugraphubuntu/.luarocks/share/lua/5.1/torch/init.lua:87>
	[C]: in function 'open'
	main_test_realsense.lua:76: in main chunk
	[C]: in function 'dofile'
	...untu/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
	[C]: at 0x00405d50

终极方案，（二）（三）以及一系列连带问题，我感觉都是hdf5版本的问题，下述方法解决。
首先删除原有的hdf5包，直接删就好
然后在之前的目录下执行以下程序：

git clone https://github.com/anibali/torch-hdf5.git
cd torch-hdf5
git checkout hdf5-1.10
luarocks make hdf5-0-0.rockspec

完毕。