-
Notifications
You must be signed in to change notification settings - Fork 6.8k
Conversation
src/operator/l2_normalization-inl.h
Outdated
"If it is set to ``channel``, the operator will compute an L2 norm for each channel in " | ||
"the multidimensional array. " | ||
"If it is set to ``spatial``, the operator will compute a cross channel norm for each " | ||
"position in the multidimensional array."); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
suggest to move the long argument doc to the description of the function
src/operator/l2_normalization.cc
Outdated
|
||
out = data / sqrt(sum(data ** 2) + eps) | ||
|
||
The parameter ``mode`` can specify the dimension along which to compute L2 norm. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
explain more about general n-d input
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
also add a particular example such as l2normalize([[1,2,3],[2,3,4]]) = [[...]]]
src/operator/make_loss.cc
Outdated
w = Variable('w') | ||
out = Activation(data = data, act_type = 'sigmoid') | ||
cross_entropy = y * log(out) + (1 - y) * log(1 - out) | ||
loss = MakeLoss(w * cross_entropy) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
change to something like:
assume out
is the predicted output and label
is the true label, then cross entropy can be defined by
cross_entropy = label * log(out) + (1 - label) * log(1 - out)
now create the loss
loss = MakeLoss(cross_entropy)
in additional, we can give a scale to loss by grad_scale=.1
also explain why Makeloss is needed, namely stop_gradient
is called. also support scale
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
.MXNET_DESCRIBE("Get output from a symbol and pass 0 gradient back") | ||
.MXNET_DESCRIBE(R"code(Stops gradient computation. | ||
|
||
Stops the accumulated gradient of the inputs from flowing through this operator |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
explain more what it means, best by examples
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hey thanks for contributions! Can you please check the comments and make the requested changes?
src/operator/activation.cc
Outdated
@@ -66,15 +66,17 @@ DMLC_REGISTER_PARAMETER(ActivationParam); | |||
|
|||
MXNET_REGISTER_OP_PROPERTY(Activation, ActivationProp) | |||
.describe(R"code(Elementwise activation function. | |||
The activation operations are applied elementwisely to each array elements. The | |||
following types are supported: | |||
The activation operations are applied elementwisely to each array elements. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi "elementwisely" occurred frequently throughout the old docs but it's not a word :). Let's say "element-wise" instead (the proper term and standard on numpy docs). Thanks!
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! fixed!
src/operator/l2_normalization.cc
Outdated
for i in 0...N | ||
out[i,:,:,...,:] = data[i,:,:,...,:] / sqrt(sum(data[i,:,:,...,:] ** 2) + eps) | ||
|
||
with ``mode`` = ``channel``, it normalize each channel in the array by its L2 norm.:: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Conjugation issue: should say "it normalizes" as above (not "it normalize")
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! fixed!
src/operator/l2_normalization.cc
Outdated
for i in 0...N | ||
out[:,i,:,...,:] = data[:,i,:,...,:] / sqrt(sum(data[:,i,:,...,:] ** 2) + eps) | ||
|
||
with ``mode`` = ``spatial``, it normalize the cross channel norm for each position |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Conjugation issue: should say "it normalizes" as above (not "it normalize")
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! fixed!
src/operator/make_loss-inl.h
Outdated
DMLC_DECLARE_FIELD(valid_thresh).set_default(0.0f) | ||
.describe("regard element valid when x > valid_thresh, this is " | ||
"used only in valid normalization mode."); | ||
.describe("Clip x to 0 when x <= valid_thresh, where x is the elements in the " |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Might be more intuitive if you could just say "clip each element in the array to 0 when it is less than valid_thresh
.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! fixed!
src/operator/activation.cc
Outdated
The activation operations are applied elementwisely to each array elements. The | ||
following types are supported: | ||
.describe(R"code(Element-wise activation function. | ||
The activation operations are applied element-wisely to each array elements. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
suggest changing "element-wisely to each array elements" to "element-wise to each array element"
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks Madan! Fixed
Madan is right. Please make this change!
…On Apr 20, 2017 5:16 PM, "Madan Jampani" ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In src/operator/activation.cc
<#5886 (comment)>:
> @@ -65,16 +65,18 @@ Operator *ActivationProp::CreateOperatorEx(Context ctx, std::vector<TShape> *in_
DMLC_REGISTER_PARAMETER(ActivationParam);
MXNET_REGISTER_OP_PROPERTY(Activation, ActivationProp)
-.describe(R"code(Elementwise activation function.
-The activation operations are applied elementwisely to each array elements. The
-following types are supported:
+.describe(R"code(Element-wise activation function.
+The activation operations are applied element-wisely to each array elements.
suggest changing "element-wisely to each array elements" to "element-wise
to each array element"
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#5886 (review)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/ACR4zodHusYy0Z2YsEuYIHpZHVOB1Zomks5rx8sagaJpZM4NBEf7>
.
|
src/operator/l2_normalization.cc
Outdated
@@ -19,8 +19,60 @@ Operator* L2NormalizationProp::CreateOperator(Context ctx) const { | |||
DMLC_REGISTER_PARAMETER(L2NormalizationParam); | |||
|
|||
MXNET_REGISTER_OP_PROPERTY(L2Normalization, L2NormalizationProp) | |||
.describe("Set the l2 norm of each instance to a constant.") | |||
.add_argument("data", "NDArray-or-Symbol", "Input data to the L2NormalizationOp.") | |||
.describe(R"code(Normalize the input array using a L2 norm. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
suggest changing "a L2 norm" to "the L2 norm"
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! Fixed
@yajiedesign can you have a look why win gpu is failed, it is caused by test_operator_gpu.test_convolution_dilated_impulse_response |
@mli some accidental errors |
* edit activation doc * doc l2_normalization * edit MakeLoss doc * edit blockgrad doc * blockgrad fileline fix * edit MakeLoss doc cont. * doc change 'tensor' to 'multidimensional array' * l2normalization doc improve * makeloss doc improve, blockgrad doc improve * fix doc in activation, l2_normalization, make_loss * fix minor grammar * use .describe to avoid build failure.
…at64 as well as operator gtest framework (#5936) * Batch Norm rewrite without mshadow as well as operator gtest framework * performance testing * lint fixes * use CUDNN for this test * remove superfluous omp define * Fix file names in comments * build, run, clean gtest works (although a test is failing) * CR comments * Adjust timing tests for more strenuous sample * Remove temp resource allocation * DeviceTensor3 added, forEachFast not yet converted * DeviceTensor3 version working * DeviceTensor3 working * . * Fix for use_global_stats * fixed bug with testing suite for double (Float64) * python unit tests working for batchnorm * python unit tests * Update documentation for mxnet.initializer.Mixed (#5937) * Update documentation for SVMOutput. (#5931) * Update documentation for SVMOutput. * Update doc for SVMOutput - fix formatting. * Adding install instruction for Ubuntu-CPU-Python (#5885) * edit ndarray API docs (#5806) * edit docs in broadcast_reduce_op * edit docs in broadcast_reduce_op * minor change * lint fix * fix * mx.nd.ones * mx.nd.repeat * mx.nd.reverse * add example in repeat * optimizer update * fix nanprod * fix optimizer_op api doc * fix reduce_op api doc * fix nd.ones api doc * mx.nd.repeat doc change * Update broadcast_reduce_op.h * Symbol docs fixes (#5930) * symbol docs minor formatting changes * deepcopy, infer_shape, infer_shape_partial docs modified * Few more small fixes * arithmetic functions fixes * some more modifications * changes after review * small change * grad function note added * More API Doc Edits (#5886) * edit activation doc * doc l2_normalization * edit MakeLoss doc * edit blockgrad doc * blockgrad fileline fix * edit MakeLoss doc cont. * doc change 'tensor' to 'multidimensional array' * l2normalization doc improve * makeloss doc improve, blockgrad doc improve * fix doc in activation, l2_normalization, make_loss * fix minor grammar * use .describe to avoid build failure. * Update documentation for mxnet.image.imdecode (#5957) * Update documentation for mxnet.image.imdecode * Update documentation for mxnet.image.imdecode (clarify that we need OpenCV and not the CV2 Python library) * Fix script by adding path to Dockerfile (#5958) * Clean install script * Add test for pip installations * Remove debug statements & comments * Make test runnable as script and from framework * Fix path to Dockerfiles * Putting failing cases at the end * Update doc for Custom operator. (#5875) * Update doc for Custom operator. * Update doc for Custom operator. * Fix formating in doc for Custom operator. * Fix formating in doc for Custom operator. * Minor change to ndarray.Custom documentation. * Minor edit in doc for Custom operator. * Minor change to doc for Custom operator. Data is 'NDArray-or-Symbol'. * Minor formatting change for Custom operator documentation. * For Custom operator doc, move example into ndarray_doc.py. * Minor change in Custom operator documentation * Improve the doc of pick + Update dmlc-core (#5946) * Add PickParam to fix the docstring and the initial value for axis * Update dmlc-core * Update dmlc-core * Image docs modified (#5973) * imageIter doc modified * edited imageiter * ADD missing Libri_sample.json, FIX minor bugs in speech_recognition example (#5962) * [KVStore] Add support for other data types (#5818) * Fix kvstore type * Fix lint * Parse inputs to DataDesc * Make module support dtype * Fix lint * Add default dtype in Comm * Fix lint * Revert rename * [cpp-package] Add C++ basic tutorial and build instruction (#5971) * Add C++ basic tutorial and build instruction * Remove binaries * Fix lint * Avoid sign-compare * Update documentation for mxnet.metric.np (#5977) * Getting rid of identity (#5935) * Activation ops (#5938) * [Ops] Add op: 'relu' * Add op: 'sigmoid' * Introduce 'kernel_launch_op' * Add tests and describe; move it to elemwise_unary_op * Fix GPU version * Convert caffe AbsVal to mx.symbol.abs in caffe converter (#5984) * Correction to LSTMCell docstring (#5986) * [Module] fix input_grads order (#5980) * fix input_grads order + update dmlc-core * set label to be optional * update env_var doc (#5964) * Adjusting make, Callback removed * batch norm gpu testing * Batch Norm rewrite without mshadow as well as operator gtest framework * performance testing * lint fixes * use CUDNN for this test * remove superfluous omp define * Fix file names in comments * build, run, clean gtest works (although a test is failing) * CR comments * Adjust timing tests for more strenuous sample * Remove temp resource allocation * rearrange source into cc and cu files * lint fixes * Trigger build * Use latest mshadow * temporarily revert channel position parameter field * Add more tests for batchnorm * Add more tests for batchnorm * test_operator_gpu working for all types * Compiles after AccReal * Compiles after AccReal * All tests working * All tests working * build, run, clean gtest works (although a test is failing) * vc++ requires explicit int type for omp for loop * Repair cpp-package * signed/unsigned fixed in cuda file * lint fixes in tests and cpp-package directories * more lint * use IsWriting() helper * Fall-through for unsupported MKL shapes/types * Fall-through for unsupported MKL shapes/types * cleaner mkl_off approach * Warning only whem MKL is requested * Warning only whem MKL is requested * lint * .. * python problem fixed * python problem fixed * Merge branch 'batchnorm' into batchnorm_pr # Conflicts: # src/operator/batch_norm.cc # src/operator/batch_norm.cu # tests/cpp/operator/batchnorm_test.cc * lint fix * lint fix * lint fix * lint fix * lint fix * Fix visual c++ compile problem * . * . * All unit tests pass again * lint fix * fix strange compile errors in CUDNN batchnorm header * FInish using flags instead of bools * lint * Fix timing pass count for forward pass * Fix R script install roxygen problem * code formatting, addition of doc strings is causing IDE to add spaces before the calls * removed commented * cr comments * Change back to compilable code * For CPU mode, store as invstd * move testing code around a little * lint fix * Use AccReal in some places to avoid fp16 problems * Fix minor invstd problem in cuda version * remove unused scale param * add permutation unit test, handle cudnn doesn't like 3D * . * lint * . * Remove mkl_off * lint fix and time cudnn when enabled
…at64 as well as operator gtest framework (apache#5936) * Batch Norm rewrite without mshadow as well as operator gtest framework * performance testing * lint fixes * use CUDNN for this test * remove superfluous omp define * Fix file names in comments * build, run, clean gtest works (although a test is failing) * CR comments * Adjust timing tests for more strenuous sample * Remove temp resource allocation * DeviceTensor3 added, forEachFast not yet converted * DeviceTensor3 version working * DeviceTensor3 working * . * Fix for use_global_stats * fixed bug with testing suite for double (Float64) * python unit tests working for batchnorm * python unit tests * Update documentation for mxnet.initializer.Mixed (apache#5937) * Update documentation for SVMOutput. (apache#5931) * Update documentation for SVMOutput. * Update doc for SVMOutput - fix formatting. * Adding install instruction for Ubuntu-CPU-Python (apache#5885) * edit ndarray API docs (apache#5806) * edit docs in broadcast_reduce_op * edit docs in broadcast_reduce_op * minor change * lint fix * fix * mx.nd.ones * mx.nd.repeat * mx.nd.reverse * add example in repeat * optimizer update * fix nanprod * fix optimizer_op api doc * fix reduce_op api doc * fix nd.ones api doc * mx.nd.repeat doc change * Update broadcast_reduce_op.h * Symbol docs fixes (apache#5930) * symbol docs minor formatting changes * deepcopy, infer_shape, infer_shape_partial docs modified * Few more small fixes * arithmetic functions fixes * some more modifications * changes after review * small change * grad function note added * More API Doc Edits (apache#5886) * edit activation doc * doc l2_normalization * edit MakeLoss doc * edit blockgrad doc * blockgrad fileline fix * edit MakeLoss doc cont. * doc change 'tensor' to 'multidimensional array' * l2normalization doc improve * makeloss doc improve, blockgrad doc improve * fix doc in activation, l2_normalization, make_loss * fix minor grammar * use .describe to avoid build failure. * Update documentation for mxnet.image.imdecode (apache#5957) * Update documentation for mxnet.image.imdecode * Update documentation for mxnet.image.imdecode (clarify that we need OpenCV and not the CV2 Python library) * Fix script by adding path to Dockerfile (apache#5958) * Clean install script * Add test for pip installations * Remove debug statements & comments * Make test runnable as script and from framework * Fix path to Dockerfiles * Putting failing cases at the end * Update doc for Custom operator. (apache#5875) * Update doc for Custom operator. * Update doc for Custom operator. * Fix formating in doc for Custom operator. * Fix formating in doc for Custom operator. * Minor change to ndarray.Custom documentation. * Minor edit in doc for Custom operator. * Minor change to doc for Custom operator. Data is 'NDArray-or-Symbol'. * Minor formatting change for Custom operator documentation. * For Custom operator doc, move example into ndarray_doc.py. * Minor change in Custom operator documentation * Improve the doc of pick + Update dmlc-core (apache#5946) * Add PickParam to fix the docstring and the initial value for axis * Update dmlc-core * Update dmlc-core * Image docs modified (apache#5973) * imageIter doc modified * edited imageiter * ADD missing Libri_sample.json, FIX minor bugs in speech_recognition example (apache#5962) * [KVStore] Add support for other data types (apache#5818) * Fix kvstore type * Fix lint * Parse inputs to DataDesc * Make module support dtype * Fix lint * Add default dtype in Comm * Fix lint * Revert rename * [cpp-package] Add C++ basic tutorial and build instruction (apache#5971) * Add C++ basic tutorial and build instruction * Remove binaries * Fix lint * Avoid sign-compare * Update documentation for mxnet.metric.np (apache#5977) * Getting rid of identity (apache#5935) * Activation ops (apache#5938) * [Ops] Add op: 'relu' * Add op: 'sigmoid' * Introduce 'kernel_launch_op' * Add tests and describe; move it to elemwise_unary_op * Fix GPU version * Convert caffe AbsVal to mx.symbol.abs in caffe converter (apache#5984) * Correction to LSTMCell docstring (apache#5986) * [Module] fix input_grads order (apache#5980) * fix input_grads order + update dmlc-core * set label to be optional * update env_var doc (apache#5964) * Adjusting make, Callback removed * batch norm gpu testing * Batch Norm rewrite without mshadow as well as operator gtest framework * performance testing * lint fixes * use CUDNN for this test * remove superfluous omp define * Fix file names in comments * build, run, clean gtest works (although a test is failing) * CR comments * Adjust timing tests for more strenuous sample * Remove temp resource allocation * rearrange source into cc and cu files * lint fixes * Trigger build * Use latest mshadow * temporarily revert channel position parameter field * Add more tests for batchnorm * Add more tests for batchnorm * test_operator_gpu working for all types * Compiles after AccReal * Compiles after AccReal * All tests working * All tests working * build, run, clean gtest works (although a test is failing) * vc++ requires explicit int type for omp for loop * Repair cpp-package * signed/unsigned fixed in cuda file * lint fixes in tests and cpp-package directories * more lint * use IsWriting() helper * Fall-through for unsupported MKL shapes/types * Fall-through for unsupported MKL shapes/types * cleaner mkl_off approach * Warning only whem MKL is requested * Warning only whem MKL is requested * lint * .. * python problem fixed * python problem fixed * Merge branch 'batchnorm' into batchnorm_pr # Conflicts: # src/operator/batch_norm.cc # src/operator/batch_norm.cu # tests/cpp/operator/batchnorm_test.cc * lint fix * lint fix * lint fix * lint fix * lint fix * Fix visual c++ compile problem * . * . * All unit tests pass again * lint fix * fix strange compile errors in CUDNN batchnorm header * FInish using flags instead of bools * lint * Fix timing pass count for forward pass * Fix R script install roxygen problem * code formatting, addition of doc strings is causing IDE to add spaces before the calls * removed commented * cr comments * Change back to compilable code * For CPU mode, store as invstd * move testing code around a little * lint fix * Use AccReal in some places to avoid fp16 problems * Fix minor invstd problem in cuda version * remove unused scale param * add permutation unit test, handle cudnn doesn't like 3D * . * lint * . * Remove mkl_off * lint fix and time cudnn when enabled
* edit activation doc * doc l2_normalization * edit MakeLoss doc * edit blockgrad doc * blockgrad fileline fix * edit MakeLoss doc cont. * doc change 'tensor' to 'multidimensional array' * l2normalization doc improve * makeloss doc improve, blockgrad doc improve * fix doc in activation, l2_normalization, make_loss * fix minor grammar * use .describe to avoid build failure.
…at64 as well as operator gtest framework (apache#5936) * Batch Norm rewrite without mshadow as well as operator gtest framework * performance testing * lint fixes * use CUDNN for this test * remove superfluous omp define * Fix file names in comments * build, run, clean gtest works (although a test is failing) * CR comments * Adjust timing tests for more strenuous sample * Remove temp resource allocation * DeviceTensor3 added, forEachFast not yet converted * DeviceTensor3 version working * DeviceTensor3 working * . * Fix for use_global_stats * fixed bug with testing suite for double (Float64) * python unit tests working for batchnorm * python unit tests * Update documentation for mxnet.initializer.Mixed (apache#5937) * Update documentation for SVMOutput. (apache#5931) * Update documentation for SVMOutput. * Update doc for SVMOutput - fix formatting. * Adding install instruction for Ubuntu-CPU-Python (apache#5885) * edit ndarray API docs (apache#5806) * edit docs in broadcast_reduce_op * edit docs in broadcast_reduce_op * minor change * lint fix * fix * mx.nd.ones * mx.nd.repeat * mx.nd.reverse * add example in repeat * optimizer update * fix nanprod * fix optimizer_op api doc * fix reduce_op api doc * fix nd.ones api doc * mx.nd.repeat doc change * Update broadcast_reduce_op.h * Symbol docs fixes (apache#5930) * symbol docs minor formatting changes * deepcopy, infer_shape, infer_shape_partial docs modified * Few more small fixes * arithmetic functions fixes * some more modifications * changes after review * small change * grad function note added * More API Doc Edits (apache#5886) * edit activation doc * doc l2_normalization * edit MakeLoss doc * edit blockgrad doc * blockgrad fileline fix * edit MakeLoss doc cont. * doc change 'tensor' to 'multidimensional array' * l2normalization doc improve * makeloss doc improve, blockgrad doc improve * fix doc in activation, l2_normalization, make_loss * fix minor grammar * use .describe to avoid build failure. * Update documentation for mxnet.image.imdecode (apache#5957) * Update documentation for mxnet.image.imdecode * Update documentation for mxnet.image.imdecode (clarify that we need OpenCV and not the CV2 Python library) * Fix script by adding path to Dockerfile (apache#5958) * Clean install script * Add test for pip installations * Remove debug statements & comments * Make test runnable as script and from framework * Fix path to Dockerfiles * Putting failing cases at the end * Update doc for Custom operator. (apache#5875) * Update doc for Custom operator. * Update doc for Custom operator. * Fix formating in doc for Custom operator. * Fix formating in doc for Custom operator. * Minor change to ndarray.Custom documentation. * Minor edit in doc for Custom operator. * Minor change to doc for Custom operator. Data is 'NDArray-or-Symbol'. * Minor formatting change for Custom operator documentation. * For Custom operator doc, move example into ndarray_doc.py. * Minor change in Custom operator documentation * Improve the doc of pick + Update dmlc-core (apache#5946) * Add PickParam to fix the docstring and the initial value for axis * Update dmlc-core * Update dmlc-core * Image docs modified (apache#5973) * imageIter doc modified * edited imageiter * ADD missing Libri_sample.json, FIX minor bugs in speech_recognition example (apache#5962) * [KVStore] Add support for other data types (apache#5818) * Fix kvstore type * Fix lint * Parse inputs to DataDesc * Make module support dtype * Fix lint * Add default dtype in Comm * Fix lint * Revert rename * [cpp-package] Add C++ basic tutorial and build instruction (apache#5971) * Add C++ basic tutorial and build instruction * Remove binaries * Fix lint * Avoid sign-compare * Update documentation for mxnet.metric.np (apache#5977) * Getting rid of identity (apache#5935) * Activation ops (apache#5938) * [Ops] Add op: 'relu' * Add op: 'sigmoid' * Introduce 'kernel_launch_op' * Add tests and describe; move it to elemwise_unary_op * Fix GPU version * Convert caffe AbsVal to mx.symbol.abs in caffe converter (apache#5984) * Correction to LSTMCell docstring (apache#5986) * [Module] fix input_grads order (apache#5980) * fix input_grads order + update dmlc-core * set label to be optional * update env_var doc (apache#5964) * Adjusting make, Callback removed * batch norm gpu testing * Batch Norm rewrite without mshadow as well as operator gtest framework * performance testing * lint fixes * use CUDNN for this test * remove superfluous omp define * Fix file names in comments * build, run, clean gtest works (although a test is failing) * CR comments * Adjust timing tests for more strenuous sample * Remove temp resource allocation * rearrange source into cc and cu files * lint fixes * Trigger build * Use latest mshadow * temporarily revert channel position parameter field * Add more tests for batchnorm * Add more tests for batchnorm * test_operator_gpu working for all types * Compiles after AccReal * Compiles after AccReal * All tests working * All tests working * build, run, clean gtest works (although a test is failing) * vc++ requires explicit int type for omp for loop * Repair cpp-package * signed/unsigned fixed in cuda file * lint fixes in tests and cpp-package directories * more lint * use IsWriting() helper * Fall-through for unsupported MKL shapes/types * Fall-through for unsupported MKL shapes/types * cleaner mkl_off approach * Warning only whem MKL is requested * Warning only whem MKL is requested * lint * .. * python problem fixed * python problem fixed * Merge branch 'batchnorm' into batchnorm_pr # Conflicts: # src/operator/batch_norm.cc # src/operator/batch_norm.cu # tests/cpp/operator/batchnorm_test.cc * lint fix * lint fix * lint fix * lint fix * lint fix * Fix visual c++ compile problem * . * . * All unit tests pass again * lint fix * fix strange compile errors in CUDNN batchnorm header * FInish using flags instead of bools * lint * Fix timing pass count for forward pass * Fix R script install roxygen problem * code formatting, addition of doc strings is causing IDE to add spaces before the calls * removed commented * cr comments * Change back to compilable code * For CPU mode, store as invstd * move testing code around a little * lint fix * Use AccReal in some places to avoid fp16 problems * Fix minor invstd problem in cuda version * remove unused scale param * add permutation unit test, handle cudnn doesn't like 3D * . * lint * . * Remove mkl_off * lint fix and time cudnn when enabled
l2_normalization
Activation
BlockGrad (same as stop_gradient)
MakeLoss
@nswamy @zackchase @Roshrini @mli @madjam @piiswrong