sandbox.cuda.dnn
– cuDNN¶
cuDNN is an NVIDIA library with functionality used by deep neural network. It provides optimized versions of some operations like the convolution. cuDNN is not currently installed with CUDA 6.5. You must download and install it yourself.
To install it, decompress the downloaded file and make the *.h
and
*.so*
files available to the compilation environment.
There are at least three possible ways of doing so:
- The easiest is to include them in your CUDA installation. Copy the
*.h
files toCUDA_ROOT/include
and the*.so*
files toCUDA_ROOT/lib64
(by default,CUDA_ROOT
is/usr/local/cuda
on Linux). - Alternatively, on Linux, you can set the environment variables
LD_LIBRARY_PATH
,LIBRARY_PATH
andCPATH
to the directory extracted from the download. If needed, separate multiple directories with:
as in thePATH
environment variable. - And as a third way, also on Linux, you can copy the
*.h
files to/usr/include
and the*.so*
files to/lib64
.
By default, Theano will detect if it can use cuDNN. If so, it will use it. If not, Theano optimizations will not introduce cuDNN ops. So Theano will still work if the user did not introduce them manually.
To get an error if Theano can not use cuDNN, use this Theano flag:
optimizer_including=cudnn
.
Note
CuDNN v2 is now released, if you used any v2 release candidate, we strongly suggest that you update it to the final version. From now on, we only support the final release.
CuDNN v2 is much faster than v1. We recommend that everybody updates to v2.
Note
Normally you should not call GPU Ops directly, but the CPU interface currently does not allow all options supported by cuDNN ops. So it is possible that you will need to call them manually.
Note
The documentation of CUDNN R1 and R2 tells that, for the following 2 operations, the reproducibility is not guaranteed: cudnnConvolutionBackwardFilter and cudnnConvolutionBackwardData. Those correspond to the gradient wrt the weights and the gradient wrt the input of the convolution. They are also used sometimes in the forward pass, when they give a speed up.
Note
There is a problem we do not understand yet when cudnn paths are used with symbolic links. So avoid using that.
Functions¶
-
theano.sandbox.cuda.dnn.
dnn_conv
(img, kerns, border_mode='valid', subsample=(1, 1), conv_mode='conv', direction_hint=None, workmem=None)¶ GPU convolution using cuDNN from NVIDIA.
The memory layout to use is ‘bc01’, that is ‘batch’, ‘channel’, ‘first dim’, ‘second dim’ in that order.
Parameters: - img – images to do the convolution over
- kerns – convolution filters
- border_mode – one of ‘valid’, ‘full’; additionally, the padding size could be directly specified by an integer or a pair of integers
- subsample – perform subsampling of the output (default: (1, 1))
- conv_mode – perform convolution (kernels flipped) or cross-correlation. One of ‘conv’, ‘cross’. (default: ‘conv’)
- direction_hint – Used by graph optimizers to change algorithm choice. By default, GpuDnnConv will be used to carry out the convolution. If border_mode is ‘valid’, subsample is (1,1) and direction_hint is ‘bprop weights’, it will use GpuDnnConvGradW. If border_mode is ‘full’, subsample is (1,1) and direction_hint is not ‘forward!’, it will use GpuDnnConvGradI. This parameter is used internally by graph optimizers and may be removed at any time without a deprecation period. You have been warned.
- workmem – Specify the amount of working memory allowed.
More memory is usually faster. One of ‘none’, ‘small’ or
‘large’. (default is None which takes its value from
config.dnn.conv.workmem
)
Warning: The cuDNN library only works with GPU that have a compute capability of 3.0 or higer. This means that older GPU will not work with this Op.
-
theano.sandbox.cuda.dnn.
dnn_pool
(img, ws, stride=(1, 1), mode='max', pad=(0, 0))¶ GPU pooling using cuDNN from NVIDIA.
The memory layout to use is ‘bc01’, that is ‘batch’, ‘channel’, ‘first dim’, ‘second dim’ in that order.
Parameters: - img – images to do the pooling over
- ws – subsampling window size
- stride – subsampling stride (default: (1, 1))
- mode – one of ‘max’, ‘average’ (default: ‘max’)
- pad – (padX, padY) padding information. padX is the size of the left and right borders, padY is the size of the top and bottom borders.
Warning: The cuDNN library only works with GPU that have a compute capability of 3.0 or higer. This means that older GPU will not work with this Op.
Note: This Op implements the ignore_border=True of max_pool_2d.
Convolution Ops¶
-
class
theano.sandbox.cuda.dnn.
GpuDnnConvDesc
(border_mode, subsample=(1, 1), conv_mode='conv')¶ This Op builds a convolution descriptor for use in the other convolution operations.
see the doc of
dnn_conv()
for a description of the parameters
-
class
theano.sandbox.cuda.dnn.
GpuDnnConv
(workmem=None, inplace=False)¶ The forward convolution.
Parameters: - image –
- kernel –
- descr – the convolution descriptor
-
static
get_out_shape
(ishape, kshape, border_mode, subsample)¶ This function computes the output shape for a convolution with the specified parameters. ishape and kshape can be symbolic or scalar.
-
class
theano.sandbox.cuda.dnn.
GpuDnnConvGradW
(inplace=False)¶ The convolution gradient with respect to the weights.
Parameters: - image –
- kernel –
- descr – the convolution descriptor
-
class
theano.sandbox.cuda.dnn.
GpuDnnConvGradI
(inplace=False)¶ The convolution gradient with respect to the inputs.
Parameters: - image –
- kernel –
- descr – the convolution descriptor
Pooling Ops¶
-
class
theano.sandbox.cuda.dnn.
GpuDnnPoolDesc
(ws=(1, 1), stride=(1, 1), mode='max', pad=(0, 0))¶ This Op builds a pooling descriptor for use in the other pooling operations.
Parameters: - ws – windows size
- stride – (dx, dy)
- mode – ‘max’ or ‘average’
- pad – (padX, padY) padding information. padX is the size of the left and right borders, padY is the size of the top and bottom borders.
-
class
theano.sandbox.cuda.dnn.
GpuDnnPool
¶ Pooling.
Parameters: - img – the image 4d tensor.
- desc – the pooling descriptor.
-
class
theano.sandbox.cuda.dnn.
GpuDnnPoolGrad
¶ The pooling gradient.
Parameters: - inp – the input of the pooling.
- out – the output of the pooling in the forward.
- inp_grad – same size as out, but is the corresponding gradient information.
- desc – The pooling descriptor.
Softmax Ops¶
-
class
theano.sandbox.cuda.dnn.
GpuDnnSoftmax
(tensor_format, algo, mode)¶ Op for the cuDNN Softmax.
Parameters: - tensor_format – Whether the data format is ‘bc01’ or ‘b01c’.
- algo – ‘fast’ or ‘accurate’ indicating whether computations should be optimized for speed or accuracy respectively.
- mode – ‘instance’ or ‘channel’ indicating whether the softmax should be computed per image across ‘c01’ or per spatial location ‘01’ per image across ‘c’.
-
class
theano.sandbox.cuda.dnn.
GpuDnnSoftmaxGrad
(tensor_format, algo, mode)¶ Op for the cuDNN SoftmaxGrad.
Parameters: - tensor_format – Whether the data format is ‘bc01’ or ‘b01c’.
- algo – ‘fast’ or ‘accurate’ indicating whether computations should be optimized for speed or accuracy respectively.
- mode – ‘instance’ or ‘channel’ indicating whether the softmax should be computed per image across ‘c01’ or per spatial location ‘01’ per image across ‘c’.