[GitHub] eric-haibin-lin commented on a change in pull request #7720: [sparse] add ftrl optimizer for sparse

[git]

eric-haibin-lin commented on a change in pull request #7720: [sparse] add ftrl 
optimizer for sparse
URL: https://github.com/apache/incubator-mxnet/pull/7720#discussion_r137450963
 
 

 ##
 File path: src/operator/optimizer_op.cc
 ##
 @@ -259,5 +260,30 @@ to be 0.9 and the learning rate :math:`\eta` to be 0.0001.
 .add_argument("delta", "NDArray-or-Symbol", "delta")
 .add_arguments(RMSPropAlexParam::__FIELDS__());
 
+NNVM_REGISTER_OP(ftrl_update)
+.describe(R"code(Update function for Ftrl optimizer.
+Referenced from *Ad Click Prediction: a View from the Trenches*, available at
+http://dl.acm.org/citation.cfm?id=2488200.
+
+.. math::
+  \\eta_{t,i} = \\frac{learningrate}{\\beta+\\sqrt{\\sum_{s=1}^tg_{s,i}^t}}
 
 Review comment:
   Could you explicitly state the sparse update rule like the one here in the 
documentation?
   
https://github.com/apache/incubator-mxnet/blob/master/src/operator/optimizer_op.cc#L159
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] ptrendx commented on issue #7778: float16 has no performance improvement

[git]

ptrendx commented on issue #7778: float16 has no performance improvement
URL: 
https://github.com/apache/incubator-mxnet/issues/7778#issuecomment-327687093
 
 
   V100 as in Volta? Then the speed you see is quite low even on fp32, so I 
would say you are limited by IO here. Default number of threads for  IO is 4 
which is way to low to saturate 4 V100 - could you test with --data-nthreads N 
option (N should be 10+ for fp16 version at least). Also, if that does not 
solve the issue, could you test single GPU numbers?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] reminisce commented on issue #7759: How to use cpu kernel of an operator (which provides only cpu implementation) in gpu context?

[git]

reminisce commented on issue #7759: How to use cpu kernel of an operator (which 
provides only cpu implementation) in gpu context?
URL: 
https://github.com/apache/incubator-mxnet/issues/7759#issuecomment-327685471
 
 
   How are the operator's forward and backward kernels implemented? By kernel, 
I mean the function block that does parallel computation such as 
[this](https://github.com/apache/incubator-mxnet/blob/master/src/operator/tensor/control_flow_op.h#L43).
 In MXNet, most of the simple operators shares the same kernels between CPU and 
GPU. That's why there is a template argument called `xpu` in Forward and 
Backward function as 
[this](https://github.com/apache/incubator-mxnet/blob/master/src/operator/tensor/control_flow_op.h#L154).
 If your cpu kernel also works for gpu and you defined Forward and Backward 
functions using `xpu` template argument, what you need to do is register the 
operator in a `.cu` file as 
[this](https://github.com/apache/incubator-mxnet/blob/master/src/operator/tensor/control_flow_op.cu#L29).
   
   Sometimes, complicated operators do not share the same kernels for cpu and 
gpu as they have different parallelization approaches and each kernel has to be 
optimized against its own device.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] javelinjs commented on a change in pull request #7782: Scala fix

[git]

javelinjs commented on a change in pull request #7782: Scala fix
URL: https://github.com/apache/incubator-mxnet/pull/7782#discussion_r137445031
 
 

 ##
 File path: scala-package/core/src/test/scala/ml/dmlc/mxnet/OperatorSuite.scala
 ##
 @@ -37,7 +37,7 @@ class OperatorSuite extends FunSuite with BeforeAndAfterAll
 exec.forward()
 val forwardOutput = exec.outputs(0)
 val forwardOutputExpected = arr.reduce(_ + _)
-assert(reldiff(forwardOutput, forwardOutputExpected) < 2e-6)
+assert(reldiff(forwardOutput, forwardOutputExpected) < 5e-5)
 
 Review comment:
   LGTM
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] kevinthesun opened a new pull request #7787: Website redesign

[git]

kevinthesun opened a new pull request #7787: Website redesign
URL: https://github.com/apache/incubator-mxnet/pull/7787
 
 
   Website UI upgrade.
   Merge after https://github.com/dmlc/web-data/pull/27
   
![web1](https://user-images.githubusercontent.com/15520525/30146014-59f5eb76-934a-11e7-9c6d-6a5a328d8e76.png)
   
![web2](https://user-images.githubusercontent.com/15520525/30146016-5b9ce1e6-934a-11e7-939c-d76dc52797a9.png)
   
   @piiswrong @lupesko @sandeep-krishnamurthy @nswamy 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] qingzhouzhen commented on issue #7786: Pvanet:Deep but Lightweight Neural Neural Networks for Real-time Object Detection

[git]

qingzhouzhen commented on issue #7786: Pvanet:Deep but Lightweight Neural 
Neural Networks for Real-time Object Detection
URL: https://github.com/apache/incubator-mxnet/pull/7786#issuecomment-327680944
 
 
   How I do:
   1 pretrain pvanet with ImageNet, it is image classification, to specify, `cd 
/path/to/example/image-calssification` `python train_imagenet.py 
--data-train=/data/ILSVRC2012_img_train.rec 
--data-val=/data/ILSVRC2012_img_val.rec --model-prefix=pvanet-models/pvanet 
--network=pvanet --gpus=0,1,2,3 --disp-batch=50 --batch-size=512 --lr=0.1 
--top-k=5` , as refered in article, reshape image to 192*192
   2 when "1" step finished, copy the model to 
/path/to/incubator-mxnet/example/rcnn/model, then `bash script/pvanet_voc07.sh 
0`
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on a change in pull request #7782: Scala fix

[git]

cjolivier01 commented on a change in pull request #7782: Scala fix
URL: https://github.com/apache/incubator-mxnet/pull/7782#discussion_r137443159
 
 

 ##
 File path: src/operator/spatial_transformer.cc
 ##
 @@ -28,7 +28,7 @@
 namespace mshadow {
 template
 bool between(DType value, int lowerBound, int upperBound) {
-  return (value >= lowerBound && value <= upperBound);
+  return (static_cast(value) >= lowerBound && static_cast(value) <= 
upperBound);
 
 Review comment:
   Also note that since the comparisons are >= and <=, this correctly handles 
the rounding 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] qingzhouzhen commented on a change in pull request #7786: Pvanet:Deep but Lightweight Neural Neural Networks for Real-time Object Detection

[git]

qingzhouzhen commented on a change in pull request #7786: Pvanet:Deep but 
Lightweight Neural Neural Networks for Real-time Object Detection
URL: https://github.com/apache/incubator-mxnet/pull/7786#discussion_r137443057
 
 

 ##
 File path: example/image-classification/symbols/pvanet.py
 ##
 @@ -0,0 +1,140 @@
+import mxnet as mx
+def get_symbol(num_classes, **kwargs):
+
+data = mx.sym.Variable(name='data')
 
 Review comment:
   pvanet is a classification net, it is a little different with it's 
counterpart as in symbol_pvanet, And secondly, I cannot upload pre-trained 
model(pvanet.py) for it is too large, thus some one else could pretrain with 
pvanet.py on ImageNet
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] chowkamlee81 closed issue #7717: Subpixel convolution(state of art) implementation rather than using Deconvolution.

[git]

chowkamlee81 closed issue #7717: Subpixel convolution(state of art) 
implementation rather than using Deconvolution.
URL: https://github.com/apache/incubator-mxnet/issues/7717
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] sxjscience commented on issue #7726: Accuracy does not increase when using distributed kvstore

[git]

sxjscience commented on issue #7726: Accuracy does not increase when using 
distributed kvstore
URL: 
https://github.com/apache/incubator-mxnet/issues/7726#issuecomment-327677715
 
 
   I'll confirm if this error occurs in the latest master today.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on a change in pull request #7786: Pvanet:Deep but Lightweight Neural Neural Networks for Real-time Object Detection

[git]

piiswrong commented on a change in pull request #7786: Pvanet:Deep but 
Lightweight Neural Neural Networks for Real-time Object Detection
URL: https://github.com/apache/incubator-mxnet/pull/7786#discussion_r137441731
 
 

 ##
 File path: example/image-classification/symbols/pvanet.py
 ##
 @@ -0,0 +1,140 @@
+import mxnet as mx
+def get_symbol(num_classes, **kwargs):
+
+data = mx.sym.Variable(name='data')
 
 Review comment:
   why put the same file at two places?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on issue #7785: How can I build a conv block using other conv's weights.

[git]

szha commented on issue #7785: How can I build a conv block using other conv's 
weights.
URL: 
https://github.com/apache/incubator-mxnet/issues/7785#issuecomment-327677122
 
 
   By reuse, do you mean you would like to tie their weights? Here's an 
[example](https://github.com/apache/incubator-mxnet/blob/master/example/gluon/word_language_model/model.py#L48-L52)
 of how to tie the weights. The weights always remain the same.
   
   I'm not familiar with torch, and I'm assuming that the AddTable is a 
declarative way of writing addition. In Gluon, this should be written in 
`hybrid_forward`. In that method, `x` is your input data, and you can use the 
layers declared in the constructor such as `self.U` by calling them with the 
input data (i.e. `output = self.U(x)`), just like calling a function in python. 
Once you finish writing the calculation of `C[i]` and `U`, Addition is then the 
simple addition of `C_out + U_out`.
   
   A great place to start learning about gluon is the interactive book. 
http://gluon.mxnet.io/. The gluon related chapters might be of most interest to 
you.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] danache opened a new issue #7785: How can I build a conv block using other conv's weights.

[git]

danache opened a new issue #7785: How can I build a conv block using other 
conv's weights.
URL: https://github.com/apache/incubator-mxnet/issues/7785
 
 
   Hi.I m trying reimplement from torch7 to mxnet.Since I am new to mxnet and I 
encounter several prombles I could not find solution in documents.
   The torch code are as follows:
   ```
   local function AttentionIter(numIn, inp , lrnSize, itersize)
   local pad = math.floor(lrnSize/2)
   local U = nnlib.SpatialConvolution(numIn,1, 3, 3, 1,1,1,1)(inp)
   local spConv = nnlib.SpatialConvolution(1,1,lrnSize,lrnSize,1,1,pad,pad)
   -- need to share the parameters and the gradParameters as well
   local spConv_clone = 
spConv:clone('weight','bias','gradWeight','gradBias')
   
   local Q={}
   local C = {}
   for i=1,itersize do
   local conv 
   local Q_tmp
   
   if i==1 then
   conv = spConv(U)
   else
   conv = spConv_clone(Q[i-1])
   end
   table.insert(C,conv)
   Q_tmp = nn.Sigmoid()(nn.CAddTable(true)({C[i], U}))
   table.insert(Q,Q_tmp)
   end
   
   local pfeat = nn.CMulTable(){inp, nn.Replicate(numIn,   2){Q[itersize]}}
   return pfeat 
   end
   ```
   
   I follow the answer in  
[#7758](https://github.com/apache/incubator-mxnet/issues/7758)
   so I init a class like this:
   ```
   class AttentionIter(gluon.HybridBlock):
   def __init__(self, numIn, inp , lrnSize, itersize, **kwargs):
   super(AttentionIter, self).__init__(**kwargs)
   self.pad = np.floor(lrnSize / 2)
   self.itersize = itersize
   self.inp = inp
   self.numIn = numIn
   
   with self.name_scope():
   self.U = gluon.nn.Conv2D(1,kernel_size=3,strides=1,padding=1)
   self.spConv = 
gluon.nn.Conv2D(1,kernel_size=lrnSize,strides=1,padding=self.pad)
   Q = []
   C = []
   for i in range(1,itersize+1):
   conv = None
   if i == 1:
   conv = self.spConv(self.U)
   else:
   conv = self.spConv(Q[i - 1])
   C.append(conv)
   q_tmp =
   
   
   
   def hybrid_forward(self, F, x):
   # F is a function space that depends on the type of x
   # If x's type is NDArray, then F will be mxnet.nd
   # If x's type is Symbol, then F will be mxnet.sym
   print('type(x): {}, F: {}'.format(
   type(x).__name__, F.__name__))
   
   return x
   ```
   but I have no idea about how to reuse parameter from a conv net just like:
   ```
local spConv_clone = spConv:clone('weight','bias','gradWeight','gradBias')
   
   ```
   besides, I am comfusing about how to add get "Q_tmp"  since I have to 
implement addtable in hybrid_forward function but Q_tmp need the results.
   Thank you !
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on issue #7710: CTC ERROR WITH CUDA ILLEGAL MEMORY ACCESS ERROR

[git]

szha commented on issue #7710: CTC ERROR WITH CUDA ILLEGAL MEMORY ACCESS ERROR
URL: 
https://github.com/apache/incubator-mxnet/issues/7710#issuecomment-327673992
 
 
   Thanks. Good to know.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] jiarenyf commented on issue #7710: CTC ERROR WITH CUDA ILLEGAL MEMORY ACCESS ERROR

[git]

jiarenyf commented on issue #7710: CTC ERROR WITH CUDA ILLEGAL MEMORY ACCESS 
ERROR
URL: 
https://github.com/apache/incubator-mxnet/issues/7710#issuecomment-327673283
 
 
   @szha 
   I found that when adding forceResize in the imageIter-kwargs, the problem 
solved ...
   And I wonder why the imageIter center-crop the image when the img-size is 
not the same with the data-shape, but not given a choice to force-reshape it ...
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] jiarenyf commented on issue #7710: CTC ERROR WITH CUDA ILLEGAL MEMORY ACCESS ERROR

[git]

jiarenyf commented on issue #7710: CTC ERROR WITH CUDA ILLEGAL MEMORY ACCESS 
ERROR
URL: 
https://github.com/apache/incubator-mxnet/issues/7710#issuecomment-327673283
 
 
   @szha 
   I found that when adding forceResize in the imageIter-kwargs, the problem 
solved ...
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on issue #7784: AssertionError: Data must be list of NDArrays

[git]

piiswrong commented on issue #7784: AssertionError: Data must be list of 
NDArrays
URL: 
https://github.com/apache/incubator-mxnet/issues/7784#issuecomment-327673173
 
 
   @zhreshold 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on a change in pull request #7782: Scala fix

[git]

szha commented on a change in pull request #7782: Scala fix
URL: https://github.com/apache/incubator-mxnet/pull/7782#discussion_r137438184
 
 

 ##
 File path: scala-package/core/src/test/scala/ml/dmlc/mxnet/OperatorSuite.scala
 ##
 @@ -37,7 +37,7 @@ class OperatorSuite extends FunSuite with BeforeAndAfterAll
 exec.forward()
 val forwardOutput = exec.outputs(0)
 val forwardOutputExpected = arr.reduce(_ + _)
-assert(reldiff(forwardOutput, forwardOutputExpected) < 2e-6)
+assert(reldiff(forwardOutput, forwardOutputExpected) < 5e-5)
 
 Review comment:
   @javelinjs 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] zydz commented on issue #6846: Check failed: err == cudaSuccess (8 vs. 0) Name: MapPlanKernel ErrStr:invalid device function

[git]

zydz commented on issue #6846: Check failed: err == cudaSuccess (8 vs. 0) Name: 
MapPlanKernel ErrStr:invalid device function
URL: 
https://github.com/apache/incubator-mxnet/issues/6846#issuecomment-327670426
 
 
   @langwei @wangmir Did you solve this issue? I met this same error. 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on issue #7783: check num_provided_arg_types and nullptr in simple_bind api for perl

[git]

eric-haibin-lin commented on issue #7783: check num_provided_arg_types and 
nullptr in simple_bind api for perl
URL: https://github.com/apache/incubator-mxnet/pull/7783#issuecomment-327666795
 
 
   The type of the pointer is const int* 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7783: check num_provided_arg_types and nullptr in simple_bind api for perl

[git]

cjolivier01 commented on issue #7783: check num_provided_arg_types and nullptr 
in simple_bind api for perl
URL: https://github.com/apache/incubator-mxnet/pull/7783#issuecomment-327657703
 
 
   Not looking at the code, but is SWIG assuming the argument is an output
   value because the parameter is not declared as const in the original source?
   
   On Wed, Sep 6, 2017 at 6:30 PM Haibin Lin  wrote:
   
   > @cjolivier01  I wanted to do the same
   > thing but the file perl-package/AI-MXNetCAPI/mxnet_wrap.cxx is generated
   > by code. @sergeykolychev  is there a
   > way?
   >
   > ?
   > You are receiving this because you were mentioned.
   > Reply to this email directly, view it on GitHub
   > 
,
   > or mute the thread
   > 

   > .
   >
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7783: check num_provided_arg_types and nullptr in simple_bind api for perl

[git]

cjolivier01 commented on issue #7783: check num_provided_arg_types and nullptr 
in simple_bind api for perl
URL: https://github.com/apache/incubator-mxnet/pull/7783#issuecomment-327656166
 
 
   I mean, zero
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] Heidisnaps opened a new issue #7784: AssertionError: Data must be list of NDArrays

[git]

Heidisnaps opened a new issue #7784: AssertionError: Data must be list of 
NDArrays
URL: https://github.com/apache/incubator-mxnet/issues/7784
 
 
   For bugs or installation issues, please provide the following information.
   The more information you provide, the more likely people will be able to 
help you.
   
   ## Environment info
   Operating System:
   Distributor ID:  Ubuntu
   Description: **Ubuntu 16.04.2 LTS**
   Release: 16.04
   Codename:xenial
   
   **gcc (Ubuntu 5.4.0-6ubuntu1~16.04.4) 5.4.0 20160609**
   Copyright (C) 2015 Free Software Foundation, Inc.
   This is free software; see the source for copying conditions.  There is NO
   warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
   
   
   
   Compiler:
   
   Package used (Python/R/Scala/Julia):
   **python**
   
   MXNet version:
   mxnet-0.11.1-py3.5.egg
   MXNet commit hash (`git rev-parse HEAD`):
   0c66eff04590c5c41a743a196283d766aab04f9b
   
   If you are using python package, please provide
   
   Python version and distribution:
   **Python 3.5.3** :: Anaconda custom (64-bit)
   
   ## Error Message:
   ~incubator-mxnet/example/ssd$ python demo.py --cpu --network resnet50 
--data-shape 512
   
   prefix 
/home/ubuntu/heidi/git_project/incubator-mxnet/example/ssd/model/ssd_resnet50_5
   [01:10:12] src/nnvm/legacy_json_util.cc:190: Loading symbol saved by 
previous version v0.10.1. Attempting to upgrade...
   [01:10:12] src/nnvm/legacy_json_util.cc:198: Symbol successfully upgraded!
   Exception in thread Thread-1:
   Traceback (most recent call last):
 File "/home/ubuntu/anaconda3/envs/mxnet_ssd/lib/python3.5/threading.py", 
line 914, in _bootstrap_inner
   self.run()
 File "/home/ubuntu/anaconda3/envs/mxnet_ssd/lib/python3.5/threading.py", 
line 862, in run
   self._target(*self._args, **self._kwargs)
 File 
"/home/ubuntu/anaconda3/envs/mxnet_ssd/lib/python3.5/site-packages/mxnet-0.11.1-py3.5.egg/mxnet/io.py",
 line 394, in prefetch_func
   self.next_batch[i] = self.iters[i].next()
 File 
"/home/ubuntu/heidi/git_project/incubator-mxnet/example/ssd/dataset/iterator.py",
 line 215, in next
   pad=self.getpad(), index=self.getindex())
 File 
"/home/ubuntu/anaconda3/envs/mxnet_ssd/lib/python3.5/site-packages/mxnet-0.11.1-py3.5.egg/mxnet/io.py",
 line 155, in __init__
   assert isinstance(data, (list, tuple)), "Data must be list of NDArrays"
   AssertionError: Data must be list of NDArrays
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin opened a new pull request #7783: check num_provided_arg_types and nullptr in simple_bind api

[git]

eric-haibin-lin opened a new pull request #7783: check num_provided_arg_types 
and nullptr in simple_bind api
URL: https://github.com/apache/incubator-mxnet/pull/7783
 
 
   In the SimpleBind CAPI, it only checks if `provided_arg_stypes == nullptr ` 
to know whether any stype is passed in. In Perl, however, the default value for 
the pointer is not nullptr but some garbage value. See the code generated for 
Perl, where arg19 and arg20 are the default value for the pointers: 
   ```
   XS(_wrap_ExecutorSimpleBind) {
 {
   SymbolHandle arg1 = (SymbolHandle) 0 ;
   ... 
   mx_uint arg18 ;
   char **arg19 = (char **) 0 ;
   int *arg20 = (int *) 0 ;
   mx_uint arg21 ;
   char **arg22 = (char **) 0 ;
   int *arg23 = (int *) 0 ;
   ... 
   mx_uint temp118 ;
   char *temp218 ;
   int temp318 ;
   unsigned int val21 ;
   int ecode21 = 0 ;
   int temp123 ;
   char *temp223 ;
   ... 
   
   {
 arg19 =   // <- not nullptr
 arg20 =   // <- not nullptr 
 arg18 = 0;
 *arg19 = NULL;
 *arg20 = 0;
  
   ```
   
   
   @cjolivier01  #7577 failed Perl test because of this. I didn't check if 
there are more cases like this where CAPI assumes the default value of a 
pointer to be nullptr. @sergeykolychev 
   
   @reminisce @piiswrong @bhavinthaker 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] rahul003 commented on a change in pull request #7421: Resolve more compile warnings

[git]

rahul003 commented on a change in pull request #7421: Resolve more compile 
warnings
URL: https://github.com/apache/incubator-mxnet/pull/7421#discussion_r137425660
 
 

 ##
 File path: src/operator/bilinear_sampler.cc
 ##
 @@ -28,7 +28,7 @@
 namespace mshadow {
 template
 bool between(DType value, int lowerBound, int upperBound) {
-  return (value >= lowerBound && value <= upperBound);
+  return (static_cast(value) >= lowerBound && static_cast(value) <= 
upperBound);
 
 Review comment:
   @tornadomeet Is DType template really required for the between functions? As 
of now it doesn't look like it. Did you use templates for the future?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on a change in pull request #7772: Use memcopy instead of assigning each individual element

[git]

szha commented on a change in pull request #7772: Use memcopy instead of 
assigning each individual element
URL: https://github.com/apache/incubator-mxnet/pull/7772#discussion_r137418825
 
 

 ##
 File path: src/operator/tensor/cast_storage-inl.h
 ##
 @@ -120,9 +119,7 @@ struct CastStorageRspDnsKernel {
 IType rid = idx[i];
 dim_t dns_offset = rid * row_length;
 dim_t rsp_offset = i * row_length;
-for (dim_t col = 0; col < row_length; col++) {
-  dns[dns_offset + col] = data[rsp_offset + col];
-}
+memcpy(dns + dns_offset, data + rsp_offset, sizeof(DType) * row_length);
 
 Review comment:
   Did a bit of searching on this and the following turns up:
   1. memcpy seems to work in 
https://stackoverflow.com/questions/10456728/is-there-an-equivalent-to-memcpy-that-works-inside-a-cuda-kernel
 (might be outdated)
   2. cudaMemcpyAsync on device is available since CUDA6, though there are 
[limitations](http://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#ixzz4rwhsHEfu)
   ```
   Notes about all memcpy/memset functions:
   Only async memcpy/set functions are supported
   Only device-to-device memcpy is permitted
   May not pass in local or shared memory pointers
   ```
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 opened a new pull request #7782: Scala fix

[git]

cjolivier01 opened a new pull request #7782: Scala fix
URL: https://github.com/apache/incubator-mxnet/pull/7782
 
 
   Adjust test threashold, which causes a test failure about 1% of the time.
   Also fix a signed/unsigned comparison warning.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on a change in pull request #7772: Use memcopy instead of assigning each individual element

[git]

eric-haibin-lin commented on a change in pull request #7772: Use memcopy 
instead of assigning each individual element
URL: https://github.com/apache/incubator-mxnet/pull/7772#discussion_r137411486
 
 

 ##
 File path: src/operator/tensor/cast_storage-inl.h
 ##
 @@ -120,9 +119,7 @@ struct CastStorageRspDnsKernel {
 IType rid = idx[i];
 dim_t dns_offset = rid * row_length;
 dim_t rsp_offset = i * row_length;
-for (dim_t col = 0; col < row_length; col++) {
-  dns[dns_offset + col] = data[rsp_offset + col];
-}
+memcpy(dns + dns_offset, data + rsp_offset, sizeof(DType) * row_length);
 
 Review comment:
   @szha  You cannot call `cudaMemcpyAsync` inside a cuda kernel. The only way 
to copy elements inside cuda kernel is to perform elemwise assignment. 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated: Fix Jenkins R gpu compilation failure (#7779)

[jxie]

This is an automated email from the ASF dual-hosted git repository.

jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git


The following commit(s) were added to refs/heads/master by this push:
 new 84f8aac  Fix Jenkins R gpu compilation failure (#7779)
84f8aac is described below

commit 84f8aac14344609c5979ebd47204162902f45b2f
Author: yuruofeifei 
AuthorDate: Wed Sep 6 15:27:08 2017 -0700

Fix Jenkins R gpu compilation failure (#7779)
---
 tests/ci_build/install/ubuntu_install_r.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tests/ci_build/install/ubuntu_install_r.sh 
b/tests/ci_build/install/ubuntu_install_r.sh
index 38d89a3..6d34441 100755
--- a/tests/ci_build/install/ubuntu_install_r.sh
+++ b/tests/ci_build/install/ubuntu_install_r.sh
@@ -24,5 +24,5 @@ gpg --keyserver keyserver.ubuntu.com --recv-key E084DAB9
 gpg -a --export E084DAB9 | apt-key add -
 
 apt-get update
-apt-get install -y r-base r-base-dev libxml2-dev libssl-dev
+apt-get install -y r-base r-base-dev libxml2-dev libssl-dev libxt-dev
 

-- 
To stop receiving notification emails like this one, please contact
['"comm...@mxnet.apache.org" '].

[GitHub] piiswrong closed pull request #7779: Fix Jenkins R gpu compilation failure

[git]

piiswrong closed pull request #7779: Fix Jenkins R gpu compilation failure
URL: https://github.com/apache/incubator-mxnet/pull/7779
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 closed issue #7675: [Jenkins CI] Broken R Unit test during build

[git]

cjolivier01 closed issue #7675: [Jenkins CI] Broken R Unit test during build
URL: https://github.com/apache/incubator-mxnet/issues/7675
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during build

[git]

cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during 
build
URL: 
https://github.com/apache/incubator-mxnet/issues/7675#issuecomment-327628757
 
 
   Issue is thought to be resolved with this change
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7686: [R] update Makefile. close #7675

[git]

cjolivier01 commented on issue #7686: [R] update Makefile. close #7675
URL: https://github.com/apache/incubator-mxnet/pull/7686#issuecomment-327628552
 
 
   Issue is thought to be resolved
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7686: [R] update Makefile. close #7675

[git]

cjolivier01 commented on issue #7686: [R] update Makefile. close #7675
URL: https://github.com/apache/incubator-mxnet/pull/7686#issuecomment-327628552
 
 
   Issue is thought to be resolved with this change
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

szha commented on a change in pull request #7727: change contrib CTC to 
zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137405194
 
 

 ##
 File path: src/operator/contrib/ctc_loss.cc
 ##
 @@ -35,7 +35,7 @@ ctcStatus_t compute_ctc_cost(const Tensor 
activations,
  void *workspace, int train) {
   int minibatch = static_cast(activations.size(1));
   int alphabet_size = static_cast(activations.size(2));
-  int blank_label = 0;
+  int blank_label = alphabet_size-1;
 
 Review comment:
   Thanks, @zhiheng-huang. I will make an example like tf doc did to clarify.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during build

[git]

cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during 
build
URL: 
https://github.com/apache/incubator-mxnet/issues/7675#issuecomment-327594119
 
 
   Which ones is known to work?
   
   On Wed, Sep 6, 2017 at 12:04 PM Qiang Kou (KK) 
   wrote:
   
   > The problem is the command executed by Jenkins server is not the same with
   > commands in Jenkinsfile.
   >
   > For example, the command executed in
   > 
https://builds.apache.org/blue/organizations/jenkins/incubator-mxnet/detail/master/287/pipeline/280
   >
   > tests/ci_build/ci_build.sh cpu R CMD INSTALL 
--library=/workspace/ut-r-cpu/site-library mxnet_current_r.tar.gz
   >
   > However, in our Jenkinsfile (
   > https://github.com/apache/incubator-mxnet/blob/master/Jenkinsfile#L382):
   >
   > sh "${docker_run} cpu R CMD INSTALL 
--library=/workspace/ut-r-cpu/site-library R-package"
   >
   > Sorry that I don't know why they are different.
   >
   > ?
   > You are receiving this because you commented.
   > Reply to this email directly, view it on GitHub
   > 
,
   > or mute the thread
   > 

   > .
   >
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during build

[git]

cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during 
build
URL: 
https://github.com/apache/incubator-mxnet/issues/7675#issuecomment-327628058
 
 
   Fixed
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cswiercz commented on issue #7781: Implement Khatri-Rao operator

[git]

cswiercz commented on issue #7781: Implement Khatri-Rao operator
URL: https://github.com/apache/incubator-mxnet/pull/7781#issuecomment-327627837
 
 
   Pinging @jli05 and @JeanKossaifi 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated: fix doc (#7780)

[jxie]

This is an automated email from the ASF dual-hosted git repository.

jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git


The following commit(s) were added to refs/heads/master by this push:
 new e242fd2  fix doc (#7780)
e242fd2 is described below

commit e242fd231e667c9d9b0c0fa439afb0c2f2cdcf81
Author: Joshua Z. Zhang 
AuthorDate: Wed Sep 6 15:18:22 2017 -0700

fix doc (#7780)
---
 docs/api/python/image.md   | 12 ++--
 src/operator/contrib/multibox_target-inl.h |  2 +-
 2 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/docs/api/python/image.md b/docs/api/python/image.md
index 23b5ee3..766a724 100644
--- a/docs/api/python/image.md
+++ b/docs/api/python/image.md
@@ -116,17 +116,17 @@ index_1  label_1  image_path_1
 Where `label_N` is a number a of fixed-width vector.
 The format of label used in object detection is a variable length vector
 ```
-A  B  [header]  [(object0), (object1), ... (objectN)]
+A  B  [extra header]  [(object0), (object1), ... (objectN)]
 ```
-Where A is the width of header, B is the width of each object.
-Header is optional and used for inserting helper information such as (width, 
height).
+Where A is the width of header (2 + length of extra header), B is the width of 
each object.
+Extra header is optional and used for inserting helper information such as 
(width, height).
 Each object is usually 5 or 6 numbers describing the object properties, for 
example:
 [id, xmin, ymin, xmax, ymax, difficulty]
 Putting all together, we have a `lst` file for object detection:
 ```
-0  2  5  640  480  1  0.1  0.2  0.8  0.9  2  0.5  0.3  0.6  0.8  data/xxx.jpg
-1  2  5  480  640  3  0.05  0.16  0.75  0.9  data/xxx.jpg
-2  2  5  500  600  2  0.6  0.1  0.7  0.5  0  0.1  0.3  0.2  0.4  3  0.25  0.25 
 0.3  0.3 data/xxx.jpg
+0  4  5  640  480  1  0.1  0.2  0.8  0.9  2  0.5  0.3  0.6  0.8  data/xxx.jpg
+1  4  5  480  640  3  0.05  0.16  0.75  0.9  data/yyy.jpg
+2  4  5  500  600  2  0.6  0.1  0.7  0.5  0  0.1  0.3  0.2  0.4  3  0.25  0.25 
 0.3  0.3 data/zzz.jpg
 ...
 ```
 
diff --git a/src/operator/contrib/multibox_target-inl.h 
b/src/operator/contrib/multibox_target-inl.h
index f76df35..872ddde 100644
--- a/src/operator/contrib/multibox_target-inl.h
+++ b/src/operator/contrib/multibox_target-inl.h
@@ -65,7 +65,7 @@ struct MultiBoxTargetParam : public 
dmlc::Parameter {
   nnvm::Tuple variances;
   DMLC_DECLARE_PARAMETER(MultiBoxTargetParam) {
 DMLC_DECLARE_FIELD(overlap_threshold).set_default(0.5f)
-.describe("Anchor-GT overlap threshold to be regarded as a possitive 
match.");
+.describe("Anchor-GT overlap threshold to be regarded as a positive 
match.");
 DMLC_DECLARE_FIELD(ignore_label).set_default(-1.0f)
 .describe("Label for ignored anchors.");
 DMLC_DECLARE_FIELD(negative_mining_ratio).set_default(-1.0f)

-- 
To stop receiving notification emails like this one, please contact
['"comm...@mxnet.apache.org" '].

[GitHub] piiswrong closed pull request #7780: fix doc

[git]

piiswrong closed pull request #7780: fix doc
URL: https://github.com/apache/incubator-mxnet/pull/7780
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cswiercz opened a new pull request #7781: Implement Khatri-Rao operator

[git]

cswiercz opened a new pull request #7781: Implement Khatri-Rao operator
URL: https://github.com/apache/incubator-mxnet/pull/7781
 
 
   Create an operator for the Khatri-Rao function implemented in #6567. The 
operator accepts a variable number of input matrix arguments as well as a flag 
indicating whether the Khatri-Rao product should be computed row-wise or 
column-wise. This option is provided because the row-wise operation is more 
memory efficient.
   
   ### A Note to Reviewers
   
   This is my first time putting MXNet code up for review so please tear this 
code apart. I need to learn.
   
   Additionally, I've knowingly omitted the Backwards/Gradient calculation.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] zhreshold opened a new pull request #7780: fix doc

[git]

zhreshold opened a new pull request #7780: fix doc
URL: https://github.com/apache/incubator-mxnet/pull/7780
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] yuruofeifei opened a new pull request #7779: Fix Jenkins R gpu compilation failure

[git]

yuruofeifei opened a new pull request #7779: Fix Jenkins R gpu compilation 
failure
URL: https://github.com/apache/incubator-mxnet/pull/7779
 
 
   xlib-backend.c:34:74: fatal error: X11/Intrinsic.h: No such file or directory
   #include  /*-> Xlib.h Xutil.h Xresource.h .. */
   ^
   compilation terminated.
   make[1]: *** [xlib-backend.o] Error 1
   make[1]: Leaving directory `/tmp/RtmpgjTXsi/R.INSTALLbad63cf92901/Cairo/src'
   ERROR: compilation failed for package 'Cairo'
   
   Fixing by installing necessary dependencies: libxt-dev
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] zhiheng-huang commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

zhiheng-huang commented on a change in pull request #7727: change contrib CTC 
to zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137400079
 
 

 ##
 File path: src/operator/contrib/ctc_loss.cc
 ##
 @@ -35,7 +35,7 @@ ctcStatus_t compute_ctc_cost(const Tensor 
activations,
  void *workspace, int train) {
   int minibatch = static_cast(activations.size(1));
   int alphabet_size = static_cast(activations.size(2));
-  int blank_label = 0;
+  int blank_label = alphabet_size-1;
 
 Review comment:
   On gluon/loss.py line 330, the new change specifies the following, which is 
inconsistent and should be corrected. Yes. It is better that alphabet_size 
includes blank label (with the id of alphabet_size-1).
   ```
   +input has shape `(sequence_length, batch_size, alphabet_size+1)`
+Note that the last dimension with index `alphabet_size` is 
reserved for special
+blank character.
   ```
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] zhiheng-huang commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

zhiheng-huang commented on a change in pull request #7727: change contrib CTC 
to zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137400079
 
 

 ##
 File path: src/operator/contrib/ctc_loss.cc
 ##
 @@ -35,7 +35,7 @@ ctcStatus_t compute_ctc_cost(const Tensor 
activations,
  void *workspace, int train) {
   int minibatch = static_cast(activations.size(1));
   int alphabet_size = static_cast(activations.size(2));
-  int blank_label = 0;
+  int blank_label = alphabet_size-1;
 
 Review comment:
   On gluon/loss.py line 330, the new change specifies the following, which is 
inconsistent and should be corrected. Yes. It is better that alphabet_size 
includes blank label (with the id of alphabet_size-1).
   
   +input has shape `(sequence_length, batch_size, alphabet_size+1)`
+Note that the last dimension with index `alphabet_size` is 
reserved for special
+blank character.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on issue #7778: float16 has no performance improvement

[git]

piiswrong commented on issue #7778: float16 has no performance improvement
URL: 
https://github.com/apache/incubator-mxnet/issues/7778#issuecomment-327621071
 
 
   @ptrendx 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] reminisce commented on issue #7774: 3d convolution gives zero output

[git]

reminisce commented on issue #7774: 3d convolution gives zero output
URL: 
https://github.com/apache/incubator-mxnet/issues/7774#issuecomment-327618553
 
 
   You need to change `args['data']=data` to `args['data'][:] = data` so that 
the content of the `data` can be copied from cpu to `args['data']` on gpu. Also 
change `conv3D_ex = conv3D.simple_bind(mx.gpu(), arg_dicts={'data':data})` to 
`conv3D_ex = conv3D.simple_bind(mx.gpu(), data=data.shape)`.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] reminisce commented on issue #7774: 3d convolution gives zero output

[git]

reminisce commented on issue #7774: 3d convolution gives zero output
URL: 
https://github.com/apache/incubator-mxnet/issues/7774#issuecomment-327618553
 
 
   You need to change `args['data']=data` to `args['data'][:] = data` so that 
the content of the `data` can be copied from cpu to `args['data']` on gpu.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong closed pull request #7776: multinode dist_sync fails with mklml

[git]

piiswrong closed pull request #7776: multinode dist_sync fails with mklml
URL: https://github.com/apache/incubator-mxnet/pull/7776
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated: multinode dist_sync fails with mklml (#7776)

[jxie]

This is an automated email from the ASF dual-hosted git repository.

jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git


The following commit(s) were added to refs/heads/master by this push:
 new 890e8f2  multinode dist_sync fails with mklml (#7776)
890e8f2 is described below

commit 890e8f2ae2b94633bbdc0b466cadce3952688576
Author: Ashok Emani 
AuthorDate: Wed Sep 6 14:19:20 2017 -0700

multinode dist_sync fails with mklml (#7776)

multinode dist_sync fails with mklml, this patch uses utility function in 
tensor_blob to handle it correctly.
---
 src/kvstore/kvstore_dist.h | 8 
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/src/kvstore/kvstore_dist.h b/src/kvstore/kvstore_dist.h
index 399754f..0cc83a0 100644
--- a/src/kvstore/kvstore_dist.h
+++ b/src/kvstore/kvstore_dist.h
@@ -134,7 +134,7 @@ class KVStoreDist : public KVStoreLocal {
 #if MKL_EXPERIMENTAL == 1
 mkl_set_tblob_eager_mode(recv_buf.data());
 #endif
-real_t* data = static_cast(recv_buf.data().dptr_);
+real_t* data = recv_buf.data().dptr();
 // false means not to delete data when SArray is deleted
 auto vals = new ps::SArray(data, size, false);
 // issue pull
@@ -297,7 +297,7 @@ class KVStoreDist : public KVStoreLocal {
 #if MKL_EXPERIMENTAL == 1
   mkl_set_tblob_eager_mode(send_buf.data());
 #endif
-  real_t* data = static_cast(send_buf.data().dptr_);
+  real_t* data = send_buf.data().dptr();
   // do push. false means no delete
   ps::SArray vals(data, size, false);
   CHECK_NOTNULL(ps_worker_)->ZPush(
@@ -330,7 +330,7 @@ class KVStoreDist : public KVStoreLocal {
 #if MKL_EXPERIMENTAL == 1
   mkl_set_tblob_eager_mode(recv_buf->data());
 #endif
-  real_t* data = static_cast(recv_buf->data().dptr_);
+  real_t* data = recv_buf->data().dptr();
   auto indices_data = indices.data();
   const auto offsets = indices_data.dptr();
   const auto unit_len = recv_buf->shape().ProdShape(1, 
recv_buf->shape().ndim());
@@ -367,7 +367,7 @@ class KVStoreDist : public KVStoreLocal {
 #if MKL_EXPERIMENTAL == 1
   mkl_set_tblob_eager_mode(send_buf.data());
 #endif
-  real_t* data = static_cast(send_buf.data().dptr_);
+  real_t* data = send_buf.data().dptr();
   bool init = send_buf.storage_initialized();
   const int64_t num_rows = init ? send_buf.aux_shape(kIdx)[0] : 0;
   const auto offsets = init ? send_buf.aux_data(kIdx).dptr() : 
nullptr;

-- 
To stop receiving notification emails like this one, please contact
['"comm...@mxnet.apache.org" '].

[GitHub] eric-haibin-lin commented on a change in pull request #7741: New operators linalg_syrk, linalg_gelqf

[git]

eric-haibin-lin commented on a change in pull request #7741: New operators 
linalg_syrk, linalg_gelqf
URL: https://github.com/apache/incubator-mxnet/pull/7741#discussion_r137392868
 
 

 ##
 File path: docs/api/python/symbol.md
 ##
 @@ -526,6 +526,8 @@ Composite multiple symbols into a new one by an operator.
 linalg_trmm
 linalg_trsm
 linalg_sumlogdiag
+linalg_syrk
+linalg_gelqf
 
 Review comment:
   I assume `ndarray.md` should be updated, too?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] asmushetzel commented on issue #7741: New operators linalg_syrk, linalg_gelqf

[git]

asmushetzel commented on issue #7741: New operators linalg_syrk, linalg_gelqf
URL: https://github.com/apache/incubator-mxnet/pull/7741#issuecomment-327610685
 
 
   Thank you Eric for the incredible turnaround time 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong closed pull request #7741: New operators linalg_syrk, linalg_gelqf

[git]

piiswrong closed pull request #7741: New operators linalg_syrk, linalg_gelqf
URL: https://github.com/apache/incubator-mxnet/pull/7741
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated: New operators linalg_syrk, linalg_gelqf. Numerical unit tests can run in float64 now (#7741)

[jxie]

This is an automated email from the ASF dual-hosted git repository.

jxie pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git


The following commit(s) were added to refs/heads/master by this push:
 new 541158d  New operators linalg_syrk, linalg_gelqf. Numerical unit tests 
can run in float64 now (#7741)
541158d is described below

commit 541158de130034d0f1b9db16b78a4bb567444f1d
Author: mseeger 
AuthorDate: Wed Sep 6 22:55:25 2017 +0200

New operators linalg_syrk, linalg_gelqf. Numerical unit tests can run in 
float64 now (#7741)
---
 docs/api/python/symbol.md  |   2 +
 python/mxnet/test_utils.py |  76 --
 src/operator/c_lapack_api.h|  66 +
 src/operator/linalg.h  |  41 +++
 src/operator/linalg_impl.h | 133 ++
 src/operator/tensor/la_op.cc   | 324 ---
 src/operator/tensor/la_op.cu   |   2 +-
 src/operator/tensor/la_op.h| 159 ++--
 src/operator/tensor/la_op_inline.h | 284 ++--
 tests/python/unittest/test_operator.py | 460 +
 10 files changed, 1217 insertions(+), 330 deletions(-)

diff --git a/docs/api/python/symbol.md b/docs/api/python/symbol.md
index d7b7359..05d9468 100644
--- a/docs/api/python/symbol.md
+++ b/docs/api/python/symbol.md
@@ -526,6 +526,8 @@ Composite multiple symbols into a new one by an operator.
 linalg_trmm
 linalg_trsm
 linalg_sumlogdiag
+linalg_syrk
+linalg_gelqf
 ```
 
 ### Miscellaneous
diff --git a/python/mxnet/test_utils.py b/python/mxnet/test_utils.py
index 4394172..946314f 100644
--- a/python/mxnet/test_utils.py
+++ b/python/mxnet/test_utils.py
@@ -447,7 +447,7 @@ def simple_forward(sym, ctx=None, is_train=False, **inputs):
 return outputs
 
 
-def _parse_location(sym, location, ctx):
+def _parse_location(sym, location, ctx, dtype=default_dtype()):
 """Parses the given location to a dictionary.
 
 Arguments of the provided op `sym` are used as dictionary keys
@@ -468,6 +468,8 @@ def _parse_location(sym, location, ctx):
 *In either case, value of all the arguments must be provided.*
 ctx : Context
 Device context.
+dtype: np.float32 or np.float64
+Datatype for mx.nd.array.
 
 Returns
 ---
@@ -489,6 +491,7 @@ def _parse_location(sym, location, ctx):
 ValueError: Symbol arguments and keys of the given location do not match.
 """
 assert isinstance(location, (dict, list, tuple))
+assert dtype == np.float32 or dtype == np.float64
 if isinstance(location, dict):
 if set(location.keys()) != set(sym.list_arguments()):
 raise ValueError("Symbol arguments and keys of the given location 
do not match."
@@ -496,12 +499,12 @@ def _parse_location(sym, location, ctx):
  % (str(set(sym.list_arguments())), 
str(set(location.keys()
 else:
 location = {k: v for k, v in zip(sym.list_arguments(), location)}
-location = {k: mx.nd.array(v, ctx=ctx) if isinstance(v, np.ndarray) \
+location = {k: mx.nd.array(v, ctx=ctx, dtype=dtype) if isinstance(v, 
np.ndarray) \
else v for k, v in location.items()}
 return location
 
 
-def _parse_aux_states(sym, aux_states, ctx):
+def _parse_aux_states(sym, aux_states, ctx, dtype=default_dtype()):
 """Parses the given auxiliary states to a dictionary.
 
 Auxiliary states of the provided op `sym` are used as dictionary
@@ -520,6 +523,10 @@ def _parse_aux_states(sym, aux_states, ctx):
 - if type is dict of str -> `np.ndarray`
 maps the name of arguments to the corresponding `np.ndarray`.
 *In either case, all aux states of `sym` must be provided.*
+ctx : Context
+Device context.
+dtype: np.float32 or np.float64
+Datatype for mx.nd.array.
 
 Returns
 ---
@@ -543,6 +550,7 @@ def _parse_aux_states(sym, aux_states, ctx):
 >>> _parse_aux_states(fc2, {'batchnorm0_moving_var': mean_states}, None)
 ValueError: Symbol aux_states names and given aux_states do not match.
 """
+assert dtype == np.float32 or dtype == np.float64
 if aux_states is not None:
 if isinstance(aux_states, dict):
 if set(aux_states.keys()) != set(sym.list_auxiliary_states()):
@@ -553,11 +561,12 @@ def _parse_aux_states(sym, aux_states, ctx):
 elif isinstance(aux_states, (list, tuple)):
 aux_names = sym.list_auxiliary_states()
 aux_states = {k:v for k, v in zip(aux_names, aux_states)}
-aux_states = {k: mx.nd.array(v, ctx=ctx) for k, v in 
aux_states.items()}
+aux_states = {k: mx.nd.array(v, ctx=ctx, dtype=dtype) for k, v in 
aux_states.items()}
 return aux_states
 
 
-def numeric_grad(executor, location, aux_states=None, eps=1e-4, 
use_forward_train=True):
+def numeric_grad(executor, location, 

[GitHub] hfutxrg opened a new issue #7778: float16 has no performance improvement

[git]

hfutxrg opened a new issue #7778: float16 has no performance improvement
URL: https://github.com/apache/incubator-mxnet/issues/7778
 
 
   For bugs or installation issues, please provide the following information.
   The more information you provide, the more likely people will be able to 
help you.
   
   ## Environment info
   Operating System: RHEL 7.3
   
   Compiler: GCC 4.8.5
   
   Package used (Python/R/Scala/Julia): Python
   
   MXNet version: 0.11.1
   
   Or if installed from source: Yes
   
   MXNet commit hash (`git rev-parse HEAD`): 
ef16cc28635d0c1ad8889232f7461c2d41fa61c7
   
   If you are using python package, please provide
   
   Python version and distribution: 2.7.5
   
   If you are using R package, please provide
   
   R `sessionInfo()`:
   
   ## Error Message:
   no error message
   
   ## Steps to reproduce
   I was training resnet50 using train_imagenet.py in 
mxnet/example/image-classification folder. I compiled mxnet with CUDA 9, CUDNN 
7.0 and run it on four V100 GPUs. 
   
   The command I used to run with the default Float32 is:
   ``python train_imagenet.py \``
   ``--network resnet \``
   ``--num-layers 50 \``
   ``--gpus 0,1,2,3 \``
   ``--num-examples 1281167 \``
  `` --num-epochs 1 \``
   ``--data-train /cm/shared/scratch/database/mxnet/ilsvrc12_train.rec \``
   ``--data-val /cm/shared/scratch/database/mxnet/ilsvrc12_val.rec \``
   ``--kv-store device``
   
   To use Float16, I just added the option:
   "--dtype float16"
   
   The training speed is 1036.15 samples/sec with float32 and 1034.27 
samples/sec for float16. So why there is no performance improvement at all? 
   
   ## What have you tried to solve it?
   I tried nvidia branch of Caffe and got close to 2x speedup with FP16 
compared to FP32. 
   
   Regards,
   Rengan
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on a change in pull request #7570: Gluon InstanceNorm and ReflectancePadding

[git]

szha commented on a change in pull request #7570: Gluon InstanceNorm and 
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#discussion_r137379884
 
 

 ##
 File path: python/mxnet/gluon/nn/basic_layers.py
 ##
 @@ -453,31 +448,65 @@ class InstanceNorm(HybridBlock):
 .. math::
 
   out = \frac{x - mean[data]}{ \sqrt{Var[data]} + \epsilon} * gamma + beta
-
+  
 
 Review comment:
   Remove the bleeding whitespace.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

szha commented on a change in pull request #7727: change contrib CTC to 
zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137379651
 
 

 ##
 File path: example/ctc/lstm_ocr.py
 ##
 @@ -59,7 +59,7 @@ def gen_rand():
 def get_label(buf):
 ret = np.zeros(4)
 for i in range(len(buf)):
-ret[i] = 1 + int(buf[i])
+ret[i] = int(buf[i])
 
 Review comment:
   With this change, alphabet_size is reserved for CTC blank label.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] zhreshold commented on a change in pull request #7626: modify gluon image_classification for imagenet

[git]

zhreshold commented on a change in pull request #7626: modify gluon 
image_classification for imagenet
URL: https://github.com/apache/incubator-mxnet/pull/7626#discussion_r135401428
 
 

 ##
 File path: src/io/iter_image_recordio_2.cc
 ##
 @@ -388,44 +406,23 @@ inline void 
ImageRecordIOParser2::ParseChunk(dmlc::InputSplit::Blob * chu
   if (!std::is_same::value) {
 // normalize/mirror here to avoid memory copies
 // logic from iter_normalize.h, function SetOutImg
-
-if (normalize_param_.mean_r > 0.0f || normalize_param_.mean_g > 
0.0f ||
-normalize_param_.mean_b > 0.0f || normalize_param_.mean_a > 
0.0f) {
-  // subtract mean per channel
-  RGBA[0] -= normalize_param_.mean_r;
-  if (n_channels >= 3) {
-RGBA[1] -= normalize_param_.mean_g;
-RGBA[2] -= normalize_param_.mean_b;
-  }
-  if (n_channels == 4) {
-RGBA[3] -= normalize_param_.mean_a;
-  }
-  for (int k = 0; k < n_channels; ++k) {
-RGBA[k] = RGBA[k] * contrast_scaled + illumination_scaled;
-  }
-} else if (!meanfile_ready_ || normalize_param_.mean_img.length() 
== 0) {
-  // do not subtract anything
-  for (int k = 0; k < n_channels; ++k) {
-RGBA[k] = RGBA[k] * normalize_param_.scale;
-  }
-} else {
-  CHECK(meanfile_ready_);
-  for (int k = 0; k < n_channels; ++k) {
-  RGBA[k] = (RGBA[k] - meanimg_[k][i][j]) * contrast_scaled + 
illumination_scaled;
+for (int k = 0; k < n_channels; ++k) {
 
 Review comment:
   The previous one is wrong, if meanfile_ is ready, the normalize_param_.scale 
was not multiplied in the previous impl
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on issue #7390: adding ranking metrics (precision/recall) at position K.

[git]

eric-haibin-lin commented on issue #7390: adding ranking metrics 
(precision/recall) at position K. 
URL: https://github.com/apache/incubator-mxnet/pull/7390#issuecomment-327600113
 
 
   what about free functions such as `true_positives`, `false_positives` and 
`false_negatives` which can be used to compute precision and recall? @nikrao 
   like the ones in tf 
https://github.com/tensorflow/tensorflow/blob/r1.3/tensorflow/python/ops/metrics_impl.py#L1565
  
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during build

[git]

cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during 
build
URL: 
https://github.com/apache/incubator-mxnet/issues/7675#issuecomment-327594119
 
 
   Which ones is known to work?
   
   On Wed, Sep 6, 2017 at 12:04 PM Qiang Kou (KK) 
   wrote:
   
   > The problem is the command executed by Jenkins server is not the same with
   > commands in Jenkinsfile.
   >
   > For example, the command executed in
   > 
https://builds.apache.org/blue/organizations/jenkins/incubator-mxnet/detail/master/287/pipeline/280
   >
   > tests/ci_build/ci_build.sh cpu R CMD INSTALL 
--library=/workspace/ut-r-cpu/site-library mxnet_current_r.tar.gz
   >
   > However, in our Jenkinsfile (
   > https://github.com/apache/incubator-mxnet/blob/master/Jenkinsfile#L382):
   >
   > sh "${docker_run} cpu R CMD INSTALL 
--library=/workspace/ut-r-cpu/site-library R-package"
   >
   > Sorry that I don't know why they are different.
   >
   > ?
   > You are receiving this because you commented.
   > Reply to this email directly, view it on GitHub
   > 
,
   > or mute the thread
   > 

   > .
   >
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[incubator-mxnet] branch master updated: Changing each task timeout from 60minutes to 120minutes (#7777)

[madjam]

This is an automated email from the ASF dual-hosted git repository.

madjam pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git


The following commit(s) were added to refs/heads/master by this push:
 new 38381c7  Changing each task timeout from 60minutes to 120minutes 
(#)
38381c7 is described below

commit 38381c74fa1b691a5d3d0ace0036614a4c564c5a
Author: mbaijal <30911248+mbai...@users.noreply.github.com>
AuthorDate: Wed Sep 6 12:50:19 2017 -0700

Changing each task timeout from 60minutes to 120minutes (#)
---
 Jenkinsfile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Jenkinsfile b/Jenkinsfile
index 5b2f819..d6111a4 100644
--- a/Jenkinsfile
+++ b/Jenkinsfile
@@ -7,7 +7,7 @@ mx_lib = 'lib/libmxnet.so, lib/libmxnet.a, dmlc-core/libdmlc.a, 
nnvm/lib/libnnvm
 // command to start a docker container
 docker_run = 'tests/ci_build/ci_build.sh'
 // timeout in minutes
-max_time = 60
+max_time = 120
 // assign any caught errors here
 err = null
 

-- 
To stop receiving notification emails like this one, please contact
['"comm...@mxnet.apache.org" '].

[GitHub] zhiheng-huang commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

zhiheng-huang commented on a change in pull request #7727: change contrib CTC 
to zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137368739
 
 

 ##
 File path: example/ctc/lstm_ocr.py
 ##
 @@ -59,7 +59,7 @@ def gen_rand():
 def get_label(buf):
 ret = np.zeros(4)
 for i in range(len(buf)):
-ret[i] = 1 + int(buf[i])
+ret[i] = int(buf[i])
 
 Review comment:
   My understanding is that zero is reserved for CTC blank label, then we shall 
not use zero as label (which is ensured by the old code).
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] zhanghang1989 commented on a change in pull request #7570: Gluon InstanceNorm and ReflectancePadding

[git]

zhanghang1989 commented on a change in pull request #7570: Gluon InstanceNorm 
and ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#discussion_r137368264
 
 

 ##
 File path: python/mxnet/gluon/nn/basic_layers.py
 ##
 @@ -442,3 +442,92 @@ def hybrid_forward(self, F, x):
 
 def __repr__(self):
 return self.__class__.__name__
+
+
+class InstanceNorm(HybridBlock):
+"""
 
 Review comment:
   Thanks! The args are added here: 
https://github.com/apache/incubator-mxnet/pull/7570/files#diff-06285baf97b6456dd23bdb99f6e239b6R461
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] gautamkmr commented on issue #7777: Changing each task timeout from 60minutes to 120minutes

[git]

gautamkmr commented on issue #: Changing each task timeout from 60minutes 
to 120minutes
URL: https://github.com/apache/incubator-mxnet/pull/#issuecomment-327587756
 
 
   LGTM
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] mbaijal opened a new pull request #7777: Changing each task timeout from 60minutes to 120minutes

[git]

mbaijal opened a new pull request #: Changing each task timeout from 
60minutes to 120minutes
URL: https://github.com/apache/incubator-mxnet/pull/
 
 
   @madjam 
   Increased the max_time from 60 mins to 120 minutes for each task in the 
Jenkinsfile since some build jobs fail due to a timeout. 
   Moving forward, The plan is to analyze the exact time it takes for each task 
in the Jenkins Job and modify these timeouts accordingly.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] thirdwing commented on issue #7675: [Jenkins CI] Broken R Unit test during build

[git]

thirdwing commented on issue #7675: [Jenkins CI] Broken R Unit test during build
URL: 
https://github.com/apache/incubator-mxnet/issues/7675#issuecomment-327582655
 
 
   The problem is the command executed by Jenkins server is not the same with 
commands in Jenkinsfile.
   
   For example, the command executed in 
https://builds.apache.org/blue/organizations/jenkins/incubator-mxnet/detail/master/287/pipeline/280
 
   ```
   tests/ci_build/ci_build.sh cpu R CMD INSTALL 
--library=/workspace/ut-r-cpu/site-library mxnet_current_r.tar.gz
   ```
   
   However, in our Jenkinsfile 
(https://github.com/apache/incubator-mxnet/blob/master/Jenkinsfile#L382):
   
   ```
   sh "${docker_run} cpu R CMD INSTALL 
--library=/workspace/ut-r-cpu/site-library R-package"
   ```
   
   Sorry that I don't know why they are different.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong closed pull request #7688: Update doc for sparse related APIs

[git]

piiswrong closed pull request #7688: Update doc for sparse related APIs
URL: https://github.com/apache/incubator-mxnet/pull/7688
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] kevinthesun commented on issue #7690: Change gluon link

[git]

kevinthesun commented on issue #7690: Change gluon link
URL: https://github.com/apache/incubator-mxnet/pull/7690#issuecomment-327577966
 
 
   @piiswrong  Fixed
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] mseeger commented on issue #7741: New operators linalg_syrk, linalg_gelqf

[git]

mseeger commented on issue #7741: New operators linalg_syrk, linalg_gelqf
URL: https://github.com/apache/incubator-mxnet/pull/7741#issuecomment-327577379
 
 
   I managed to run all unit tests in continuous-integration/jenkins/pr-head, 
except for the R ones. They fail for funny reasons, which seem nothing to do 
with my changes.
   
   I am trying again.
   
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] nswamy closed pull request #7746: CoremlBugFixes:Input variable name can be something other than data; ?

[git]

nswamy closed pull request #7746: CoremlBugFixes:Input variable name can be 
something other than data; ?
URL: https://github.com/apache/incubator-mxnet/pull/7746
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on a change in pull request #7688: Update doc for sparse related APIs

[git]

eric-haibin-lin commented on a change in pull request #7688: Update doc for 
sparse related APIs
URL: https://github.com/apache/incubator-mxnet/pull/7688#discussion_r137350322
 
 

 ##
 File path: python/mxnet/kvstore.py
 ##
 @@ -266,8 +283,8 @@ def pull(self, key, out=None, priority=0):
 self.handle, mx_uint(len(ckeys)), ckeys, cvals, 
ctypes.c_int(priority)))
 
 def row_sparse_pull(self, key, out=None, priority=0, row_ids=None):
-""" Pulls a single row_sparse value or a sequence of row_sparse values 
from the store
- with specified row_ids.
+""" Pulls a single RowSparseNDArray value or a sequence of 
RowSparseNDArray values \
 
 Review comment:
   causes display problem at API doc. Otherwise, it doesn't treat the second 
line as part of the first line
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on a change in pull request #7772: Use memcopy instead of assigning each individual element

[git]

szha commented on a change in pull request #7772: Use memcopy instead of 
assigning each individual element
URL: https://github.com/apache/incubator-mxnet/pull/7772#discussion_r137348546
 
 

 ##
 File path: src/operator/tensor/cast_storage-inl.h
 ##
 @@ -120,9 +119,7 @@ struct CastStorageRspDnsKernel {
 IType rid = idx[i];
 dim_t dns_offset = rid * row_length;
 dim_t rsp_offset = i * row_length;
-for (dim_t col = 0; col < row_length; col++) {
-  dns[dns_offset + col] = data[rsp_offset + col];
-}
+memcpy(dns + dns_offset, data + rsp_offset, sizeof(DType) * row_length);
 
 Review comment:
   Maybe do
   ```c++
   #if defined(__CUDACC__)
   cudaMemcpyAsync(...);
   #else
   memcpy(...);
   #endif
   ```
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] reminisce commented on a change in pull request #7688: Update doc for sparse related APIs

[git]

reminisce commented on a change in pull request #7688: Update doc for sparse 
related APIs
URL: https://github.com/apache/incubator-mxnet/pull/7688#discussion_r137346068
 
 

 ##
 File path: python/mxnet/kvstore.py
 ##
 @@ -266,8 +283,8 @@ def pull(self, key, out=None, priority=0):
 self.handle, mx_uint(len(ckeys)), ckeys, cvals, 
ctypes.c_int(priority)))
 
 def row_sparse_pull(self, key, out=None, priority=0, row_ids=None):
-""" Pulls a single row_sparse value or a sequence of row_sparse values 
from the store
- with specified row_ids.
+""" Pulls a single RowSparseNDArray value or a sequence of 
RowSparseNDArray values \
 
 Review comment:
   Why need "\" at the end of the line?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

szha commented on a change in pull request #7727: change contrib CTC to 
zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137346338
 
 

 ##
 File path: example/ctc/lstm_ocr.py
 ##
 @@ -59,7 +59,7 @@ def gen_rand():
 def get_label(buf):
 ret = np.zeros(4)
 for i in range(len(buf)):
-ret[i] = 1 + int(buf[i])
+ret[i] = int(buf[i])
 
 Review comment:
   Making a 1-indexed API into 0-indexed API is not really backward compatible. 
Contrib package doesn't guarantee backward compatibility anyway.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

szha commented on a change in pull request #7727: change contrib CTC to 
zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137346031
 
 

 ##
 File path: src/operator/contrib/ctc_loss.cc
 ##
 @@ -35,7 +35,7 @@ ctcStatus_t compute_ctc_cost(const Tensor 
activations,
  void *workspace, int train) {
   int minibatch = static_cast(activations.size(1));
   int alphabet_size = static_cast(activations.size(2));
-  int blank_label = 0;
+  int blank_label = alphabet_size-1;
 
 Review comment:
   Yes. Alphabet size may be a misnomer, because the dimension of data is 
*alphabet_size + 1*, and this dimension is passed in as `alphabet_size`
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on issue #7712: Python symbol/ndarray api re-org

[git]

eric-haibin-lin commented on issue #7712: Python symbol/ndarray api re-org
URL: https://github.com/apache/incubator-mxnet/pull/7712#issuecomment-327564178
 
 
   @cjolivier01 @anirudh2290 @bhavinthaker 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on issue #7744: multinode dist_sync fails with mklml

[git]

piiswrong commented on issue #7744: multinode dist_sync fails with mklml
URL: https://github.com/apache/incubator-mxnet/pull/7744#issuecomment-327563518
 
 
   Please submit your change to master. We will decided whether to merge it to 
0.11 later
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on issue #7712: Python symbol/ndarray api re-org

[git]

eric-haibin-lin commented on issue #7712: Python symbol/ndarray api re-org
URL: https://github.com/apache/incubator-mxnet/pull/7712#issuecomment-327563726
 
 
   Done. Also ping @reminisce 
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on a change in pull request #7570: Gluon InstanceNorm and ReflectancePadding

[git]

piiswrong commented on a change in pull request #7570: Gluon InstanceNorm and 
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#discussion_r137341446
 
 

 ##
 File path: python/mxnet/gluon/nn/basic_layers.py
 ##
 @@ -442,3 +442,92 @@ def hybrid_forward(self, F, x):
 
 def __repr__(self):
 return self.__class__.__name__
+
+
+class InstanceNorm(HybridBlock):
+"""
+Applies instance normalization to the n-dimensional input array.
+This operator takes an n-dimensional input array where (n>2) and 
normalizes 
+the input using the following formula:
+
+.. math::
+
+  out = \frac{x - mean[data]}{ \sqrt{Var[data]} + \epsilon} * gamma + beta
+
+This layer is similar to batch normalization layer (`BatchNorm`)
+with two differences: first, the normalization is
+carried out per example (instance), not over a batch. Second, the
+same normalization is applied both at test and train time. This
+operation is also known as `contrast normalization`.
+If the input data is of shape [batch, channel, spacial_dim1, spacial_dim2, 
...],
+`gamma` and `beta` parameters must be vectors of shape [channel].
+This implementation is based on paper:
+.. [1] Instance Normalization: The Missing Ingredient for Fast Stylization,
+   D. Ulyanov, A. Vedaldi, V. Lempitsky, 2016 (arXiv:1607.08022v2).
+Examples::
+
+  // Input of shape (2,1,2)
+  x = mx.nd.array([[[ 1.1,  2.2]],
+  [[ 3.3,  4.4]]])
+  // Instance normalization is calculated with the above formula
+  layer = InstanceNorm()
+  layer.initialize(ctx=mx.cpu(0))
+  layer(x)
+  [[[-0.8355  0.8331]]
+   [[-0.8319  0.8361]]]
+  
+"""
+def __init__(self, axis=1, momentum=0.9, epsilon=1e-5, center=True, 
scale=False,
+ beta_initializer='zeros', gamma_initializer='ones',
+ in_channels=0, **kwargs):
+super(InstanceNorm, self).__init__(**kwargs)
+self._kwargs = {'eps': epsilon}
+if in_channels != 0:
+self.in_channels = in_channels
+self.gamma = self.params.get('gamma', grad_req='write' if scale else 
'null',
+ shape=(in_channels,), 
init=gamma_initializer,
+ allow_deferred_init=True)
+self.beta = self.params.get('beta', grad_req='write' if center else 
'null',
+shape=(in_channels,), 
init=beta_initializer,
+allow_deferred_init=True)
+
+def hybrid_forward(self, F, x, gamma, beta):
+return F.InstanceNorm(x, gamma, beta,
+   name='fwd', **self._kwargs)
+
+def __repr__(self):
+s = '{name}({content}'
+if hasattr(self, 'in_channels'):
+s += ', in_channels={0}'.format(self.in_channels)
+s += ')'
+return s.format(name=self.__class__.__name__,
+content=', '.join(['='.join([k, v.__repr__()])
+   for k, v in self._kwargs.items()]))
+
+
+class ReflectionPad(HybridBlock):
+"""Pads the input tensor using the reflection of the input boundary.
+
+Args:
+pad_width int: the size of the padding. If is int, uses the same
+padding in all boundaries. 
+
+Shape:
+- Input: :math:`(N, C, H_{in}, W_{in})`
+- Output: :math:`(N, C, H_{out}, W_{out})` where
+  :math:`H_{out} = H_{in} + paddingTop + paddingBottom`
+  :math:`W_{out} = W_{in} + paddingLeft + paddingRight`
+
+Examples::
+
+m = nn.ReflectionPad(3)
+input = mx.nd.random_normal(shape=(16, 3, 224, 224))
+output = m(input)
+"""
+def __init__(self, pad_width=None, **kwargs):
 
 Review comment:
   what does a None default mean?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on a change in pull request #7570: Gluon InstanceNorm and ReflectancePadding

[git]

piiswrong commented on a change in pull request #7570: Gluon InstanceNorm and 
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#discussion_r137341278
 
 

 ##
 File path: python/mxnet/gluon/nn/basic_layers.py
 ##
 @@ -442,3 +442,92 @@ def hybrid_forward(self, F, x):
 
 def __repr__(self):
 return self.__class__.__name__
+
+
+class InstanceNorm(HybridBlock):
+"""
+Applies instance normalization to the n-dimensional input array.
+This operator takes an n-dimensional input array where (n>2) and 
normalizes 
+the input using the following formula:
+
+.. math::
+
+  out = \frac{x - mean[data]}{ \sqrt{Var[data]} + \epsilon} * gamma + beta
+
+This layer is similar to batch normalization layer (`BatchNorm`)
+with two differences: first, the normalization is
+carried out per example (instance), not over a batch. Second, the
+same normalization is applied both at test and train time. This
+operation is also known as `contrast normalization`.
+If the input data is of shape [batch, channel, spacial_dim1, spacial_dim2, 
...],
+`gamma` and `beta` parameters must be vectors of shape [channel].
+This implementation is based on paper:
+.. [1] Instance Normalization: The Missing Ingredient for Fast Stylization,
+   D. Ulyanov, A. Vedaldi, V. Lempitsky, 2016 (arXiv:1607.08022v2).
+Examples::
+
+  // Input of shape (2,1,2)
+  x = mx.nd.array([[[ 1.1,  2.2]],
+  [[ 3.3,  4.4]]])
+  // Instance normalization is calculated with the above formula
+  layer = InstanceNorm()
+  layer.initialize(ctx=mx.cpu(0))
+  layer(x)
+  [[[-0.8355  0.8331]]
+   [[-0.8319  0.8361]]]
+  
+"""
+def __init__(self, axis=1, momentum=0.9, epsilon=1e-5, center=True, 
scale=False,
+ beta_initializer='zeros', gamma_initializer='ones',
+ in_channels=0, **kwargs):
+super(InstanceNorm, self).__init__(**kwargs)
+self._kwargs = {'eps': epsilon}
+if in_channels != 0:
+self.in_channels = in_channels
+self.gamma = self.params.get('gamma', grad_req='write' if scale else 
'null',
+ shape=(in_channels,), 
init=gamma_initializer,
+ allow_deferred_init=True)
+self.beta = self.params.get('beta', grad_req='write' if center else 
'null',
+shape=(in_channels,), 
init=beta_initializer,
+allow_deferred_init=True)
+
+def hybrid_forward(self, F, x, gamma, beta):
+return F.InstanceNorm(x, gamma, beta,
+   name='fwd', **self._kwargs)
+
+def __repr__(self):
+s = '{name}({content}'
+if hasattr(self, 'in_channels'):
+s += ', in_channels={0}'.format(self.in_channels)
+s += ')'
+return s.format(name=self.__class__.__name__,
+content=', '.join(['='.join([k, v.__repr__()])
+   for k, v in self._kwargs.items()]))
+
+
+class ReflectionPad(HybridBlock):
 
 Review comment:
   Should be named ReflectionPad2D and put in convolution layers
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on a change in pull request #7570: Gluon InstanceNorm and ReflectancePadding

[git]

piiswrong commented on a change in pull request #7570: Gluon InstanceNorm and 
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#discussion_r137341129
 
 

 ##
 File path: python/mxnet/gluon/nn/basic_layers.py
 ##
 @@ -442,3 +442,92 @@ def hybrid_forward(self, F, x):
 
 def __repr__(self):
 return self.__class__.__name__
+
+
+class InstanceNorm(HybridBlock):
+"""
+Applies instance normalization to the n-dimensional input array.
+This operator takes an n-dimensional input array where (n>2) and 
normalizes 
+the input using the following formula:
+
+.. math::
+
+  out = \frac{x - mean[data]}{ \sqrt{Var[data]} + \epsilon} * gamma + beta
+
+This layer is similar to batch normalization layer (`BatchNorm`)
+with two differences: first, the normalization is
+carried out per example (instance), not over a batch. Second, the
+same normalization is applied both at test and train time. This
+operation is also known as `contrast normalization`.
+If the input data is of shape [batch, channel, spacial_dim1, spacial_dim2, 
...],
+`gamma` and `beta` parameters must be vectors of shape [channel].
+This implementation is based on paper:
+.. [1] Instance Normalization: The Missing Ingredient for Fast Stylization,
+   D. Ulyanov, A. Vedaldi, V. Lempitsky, 2016 (arXiv:1607.08022v2).
+Examples::
+
+  // Input of shape (2,1,2)
+  x = mx.nd.array([[[ 1.1,  2.2]],
+  [[ 3.3,  4.4]]])
+  // Instance normalization is calculated with the above formula
+  layer = InstanceNorm()
+  layer.initialize(ctx=mx.cpu(0))
+  layer(x)
+  [[[-0.8355  0.8331]]
+   [[-0.8319  0.8361]]]
+  
+"""
+def __init__(self, axis=1, momentum=0.9, epsilon=1e-5, center=True, 
scale=False,
+ beta_initializer='zeros', gamma_initializer='ones',
+ in_channels=0, **kwargs):
+super(InstanceNorm, self).__init__(**kwargs)
+self._kwargs = {'eps': epsilon}
+if in_channels != 0:
+self.in_channels = in_channels
+self.gamma = self.params.get('gamma', grad_req='write' if scale else 
'null',
+ shape=(in_channels,), 
init=gamma_initializer,
+ allow_deferred_init=True)
+self.beta = self.params.get('beta', grad_req='write' if center else 
'null',
+shape=(in_channels,), 
init=beta_initializer,
+allow_deferred_init=True)
+
+def hybrid_forward(self, F, x, gamma, beta):
+return F.InstanceNorm(x, gamma, beta,
+   name='fwd', **self._kwargs)
+
+def __repr__(self):
+s = '{name}({content}'
+if hasattr(self, 'in_channels'):
+s += ', in_channels={0}'.format(self.in_channels)
+s += ')'
+return s.format(name=self.__class__.__name__,
+content=', '.join(['='.join([k, v.__repr__()])
+   for k, v in self._kwargs.items()]))
+
+
+class ReflectionPad(HybridBlock):
+"""Pads the input tensor using the reflection of the input boundary.
+
+Args:
 
 Review comment:
   See other APIs for formatting.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on a change in pull request #7570: Gluon InstanceNorm and ReflectancePadding

[git]

piiswrong commented on a change in pull request #7570: Gluon InstanceNorm and 
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#discussion_r137341055
 
 

 ##
 File path: python/mxnet/gluon/nn/basic_layers.py
 ##
 @@ -442,3 +442,92 @@ def hybrid_forward(self, F, x):
 
 def __repr__(self):
 return self.__class__.__name__
+
+
+class InstanceNorm(HybridBlock):
+"""
 
 Review comment:
   missing argument documentation
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on issue #7570: Gluon InstanceNorm and ReflectancePadding

[git]

piiswrong commented on issue #7570: Gluon InstanceNorm and ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#issuecomment-327561522
 
 
   Front end interface APIs need to conform to the highest standards.
   
   Please look at other APIs and fix the formatting.
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on issue #7702: fix caffe model convert leakyrelu

[git]

piiswrong commented on issue #7702: fix caffe model convert leakyrelu
URL: https://github.com/apache/incubator-mxnet/pull/7702#issuecomment-327560106
 
 
   Cool. Thanks
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong closed pull request #7702: fix caffe model convert leakyrelu

[git]

piiswrong closed pull request #7702: fix caffe model convert leakyrelu
URL: https://github.com/apache/incubator-mxnet/pull/7702
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] zhanghang1989 commented on issue #7570: Gluon InstanceNorm and ReflectancePadding

[git]

zhanghang1989 commented on issue #7570: Gluon InstanceNorm and 
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#issuecomment-327560038
 
 
   Let me know if it is ready to merge. I will update another PR accordingly 
https://github.com/apache/incubator-mxnet/pull/7579. @piiswrong @szha  Thanks!
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

piiswrong commented on a change in pull request #7727: change contrib CTC to 
zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137339240
 
 

 ##
 File path: src/operator/contrib/ctc_loss.cc
 ##
 @@ -35,7 +35,7 @@ ctcStatus_t compute_ctc_cost(const Tensor 
activations,
  void *workspace, int train) {
   int minibatch = static_cast(activations.size(1));
   int alphabet_size = static_cast(activations.size(2));
-  int blank_label = 0;
+  int blank_label = alphabet_size-1;
 
 Review comment:
   Are you sure? Shouldn't it be alphabet_size?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on a change in pull request #7727: change contrib CTC to zero-indexed, in log space. sequence mask for grad

[git]

piiswrong commented on a change in pull request #7727: change contrib CTC to 
zero-indexed, in log space. sequence mask for grad
URL: https://github.com/apache/incubator-mxnet/pull/7727#discussion_r137338691
 
 

 ##
 File path: example/ctc/lstm_ocr.py
 ##
 @@ -59,7 +59,7 @@ def gen_rand():
 def get_label(buf):
 ret = np.zeros(4)
 for i in range(len(buf)):
-ret[i] = 1 + int(buf[i])
+ret[i] = int(buf[i])
 
 Review comment:
   contrib.ctcloss has been there for a while. We should maintain backward 
compatibility
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on a change in pull request #7741: New operators linalg_syrk, linalg_gelqf

[git]

piiswrong commented on a change in pull request #7741: New operators 
linalg_syrk, linalg_gelqf
URL: https://github.com/apache/incubator-mxnet/pull/7741#discussion_r137337294
 
 

 ##
 File path: docs/api/python/symbol.md
 ##
 @@ -526,6 +526,8 @@ Composite multiple symbols into a new one by an operator.
 linalg_trmm
 linalg_trsm
 linalg_sumlogdiag
+   linalg_syrk
 
 Review comment:
   wrong indentation
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] piiswrong commented on issue #7763: iOS sample app for the Core ML converter tool

[git]

piiswrong commented on issue #7763: iOS sample app for the Core ML converter 
tool
URL: https://github.com/apache/incubator-mxnet/pull/7763#issuecomment-327557536
 
 
   we don't commit binaries and images to mxnet repo. Please submit them to 
dmlc/web-data
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on issue #7720: [sparse] add ftrl optimizer for sparse

[git]

eric-haibin-lin commented on issue #7720: [sparse] add ftrl optimizer for sparse
URL: https://github.com/apache/incubator-mxnet/pull/7720#issuecomment-327552883
 
 
   Thanks for the contribution! I'll take a look after completing some 
documentation work.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on issue #7770: MKL ML results depending on number of threads

[git]

szha commented on issue #7770: MKL ML results depending on number of threads
URL: 
https://github.com/apache/incubator-mxnet/issues/7770#issuecomment-327552409
 
 
   Would you be able to isolate where the training diverges depending on the 
number of threads? @ykim362 what's MKLML's plan and timeline?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] szha commented on issue #6778: undefined symbol: dnnBatchNormalizationCreateBackward_v2_F32

[git]

szha commented on issue #6778: undefined symbol: 
dnnBatchNormalizationCreateBackward_v2_F32
URL: 
https://github.com/apache/incubator-mxnet/issues/6778#issuecomment-327551791
 
 
   @morusu which version, and which OS are you using?
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on a change in pull request #7712: python symbol/ndarray api re-org

[git]

eric-haibin-lin commented on a change in pull request #7712: python 
symbol/ndarray api re-org
URL: https://github.com/apache/incubator-mxnet/pull/7712#discussion_r137314034
 
 

 ##
 File path: docs/api/python/ndarray/ndarray.md
 ##
 @@ -625,16 +571,54 @@ We summarize the interface for each class in the 
following sections.
 Custom
 ```
 
+## Linear Algebra
 
 Review comment:
   ok.. I misunderstood your previous comment.
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during build

[git]

cjolivier01 commented on issue #7675: [Jenkins CI] Broken R Unit test during 
build
URL: 
https://github.com/apache/incubator-mxnet/issues/7675#issuecomment-327304053
 
 
   I think I know what's causing this...
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] cjolivier01 closed pull request #7747: Fix R test problem

[git]

cjolivier01 closed pull request #7747: Fix R test problem
URL: https://github.com/apache/incubator-mxnet/pull/7747
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] eric-haibin-lin commented on a change in pull request #7772: Use memcopy instead of assigning each individual element

[git]

eric-haibin-lin commented on a change in pull request #7772: Use memcopy 
instead of assigning each individual element
URL: https://github.com/apache/incubator-mxnet/pull/7772#discussion_r137280575
 
 

 ##
 File path: src/operator/tensor/cast_storage-inl.h
 ##
 @@ -120,9 +119,7 @@ struct CastStorageRspDnsKernel {
 IType rid = idx[i];
 dim_t dns_offset = rid * row_length;
 dim_t rsp_offset = i * row_length;
-for (dim_t col = 0; col < row_length; col++) {
-  dns[dns_offset + col] = data[rsp_offset + col];
-}
+memcpy(dns + dns_offset, data + rsp_offset, sizeof(DType) * row_length);
 
 Review comment:
   Thanks for the contribution. I don't think this will work on GPU though as 
this kernel is shared by CPU and GPU..
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] xuerenlv closed pull request #7775: Merge pull request #1 from apache/master

[git]

xuerenlv closed pull request #7775: Merge pull request #1 from apache/master
URL: https://github.com/apache/incubator-mxnet/pull/7775
 
 
   
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] xuerenlv opened a new pull request #7775: Merge pull request #1 from apache/master

[git]

xuerenlv opened a new pull request #7775: Merge pull request #1 from 
apache/master
URL: https://github.com/apache/incubator-mxnet/pull/7775
 
 
   update
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

[GitHub] mafhha opened a new issue #7774: 3d convolution gives zero output

[git]

mafhha opened a new issue #7774: 3d convolution gives zero output
URL: https://github.com/apache/incubator-mxnet/issues/7774
 
 
   When using mxnet 3d Convolution layer, the output is all zeros. 
   
   
   ```
   import mxnet as mx
   data = mx.nd.arange(24).reshape((2,1,2,2,3))
   
   x = mx.sym.Variable('data', shape=(2,1,2,2,3))
   conv3D = mx.sym.Convolution(data=x, kernel=(2, 2, 3), pad=(1, 1, 1), 
num_filter=2)
   
   conv3D_ex = conv3D.simple_bind(mx.gpu(), arg_dicts={'data':data})
   args = conv3D_ex.arg_dict
   
   
   init = mx.init.Uniform(scale=0.01)
   for name, arr in args.items():
   if name not in ['data']:
   init(name, arr)
   
   args['data']=data
   
   print args
   print conv3D_ex.forward()
   
   {'data': 
   [  0.   1.   2.]
   [  3.   4.   5.]]
   
  [[  6.   7.   8.]
   [  9.  10.  11.
   
   
   
 12.  13.  14.]
   [ 15.  16.  17.]]
   
  [[ 18.  19.  20.]
   [ 21.  22.  23.]
   , 'convolution0_weight': 
   [ 0.00480439  0.00841988 -0.00921959]
   [ 0.00937926  0.00850281 -0.001073  ]]
   
  [[ 0.00334638 -0.00780139 -0.00059563]
   [ 0.00026387  0.0055235  -0.00410469
   
   
   
 0.00428019 -0.00282996  0.00362837]
   [-0.0041597  -0.00361182  0.00621827]]
   
  [[-0.00691776 -0.00109668 -0.00584008]
   [ 0.00221961 -0.00385449 -0.00168849]
   , 'convolution0_bias': 
   [ 0.  0.]
   }
   [
   [ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]
   
  [[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]
   
  [[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]]
   
   
 [[[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]
   
  [[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]
   
  [[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.
   
   
   
 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]
   
  [[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]
   
  [[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]]
   
   
 [[[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]
   
  [[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]]
   
  [[ 0.  0.  0.]
   [ 0.  0.  0.]
   [ 0.  0.  0.]
   ]
   ```
 

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services