[GitHub] ashokei commented on a change in pull request #10736: [MXNET-367] update mkldnn to v0.14 and disable building test examples

[GitBox]

ashokei commented on a change in pull request #10736: [MXNET-367] update mkldnn 
to v0.14 and disable building test examples
URL: https://github.com/apache/incubator-mxnet/pull/10736#discussion_r185701283
 
 

 ##
 File path: ci/docker/install/ubuntu_mklml.sh
 ##
 @@ -21,5 +21,5 @@
 # the whole docker cache for the image
 
 set -ex
-wget --no-check-certificate -O /tmp/mklml.tgz 
https://github.com/intel/mkl-dnn/releases/download/v0.12/mklml_lnx_2018.0.1.20171227.tgz
-tar -zxvf /tmp/mklml.tgz && cp -rf mklml_*/* /usr/local/ && rm -rf mklml_*
\ No newline at end of file
+wget --no-check-certificate -O /tmp/mklml.tgz 
https://github.com/intel/mkl-dnn/releases/download/v0.14/mklml_lnx_2018.0.3.20180406.tgz
 
 Review comment:
   i notice these scripts are called only for docker, as part of docker image 
build; i think we do not have 3rdparty/mkldnn folder yet , as the actual mxnet 
build happens 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] hongtao12310 commented on issue #10579: NNPACK can't be compiled at 1.1.0 1.2.0 and master branch

[GitBox]

hongtao12310 commented on issue #10579: NNPACK can't be compiled at 1.1.0 1.2.0 
and master branch
URL: 
https://github.com/apache/incubator-mxnet/issues/10579#issuecomment-386197197
 
 
   agree @fullfanta 
   
   Let's wait for the fix


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] apshreyans commented on issue #9602: MXNET not working in R

[GitBox]

apshreyans commented on issue #9602: MXNET not working in R
URL: 
https://github.com/apache/incubator-mxnet/issues/9602#issuecomment-386197156
 
 
   Thanks much. It worked for me.
   
   On Thu, May 3, 2018, 10:39 AM jeremiedb  wrote:
   
   > Hi, the official repo hasn't been updated. In the meantime, you can
   > install this version (CPU only for now):
   >
   > 
install.packages("https://github.com/jeremiedb/mxnet_winbin/raw/master/mxnet.zip;,
 repos = NULL)
   >
   > —
   > 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] ashokei commented on a change in pull request #10736: [MXNET-367] update mkldnn to v0.14 and disable building test examples

[GitBox]

ashokei commented on a change in pull request #10736: [MXNET-367] update mkldnn 
to v0.14 and disable building test examples
URL: https://github.com/apache/incubator-mxnet/pull/10736#discussion_r185701037
 
 

 ##
 File path: prepare_mkl.sh
 ##
 @@ -58,16 +58,16 @@ MXNET_ROOT=`dirname $0`
 USE_MKLML=0
 # NOTE: if you update the following line, please also update the dockerfile at
 # tests/ci_build/Dockerfile.mkl
-VERSION_MATCH=20171227
+VERSION_MATCH=20180406
 
 Review comment:
   `USE_MKLML` is used inside mshadow build. USE_BLAS can be set to `mkl`.


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] Godricly commented on issue #9704: Misleading Document for conv1d and max1d in gluon

[GitBox]

Godricly commented on issue #9704: Misleading Document for conv1d and max1d in 
gluon
URL: 
https://github.com/apache/incubator-mxnet/issues/9704#issuecomment-386196628
 
 
   Some updates here. Unlike the pooling layers which has an 
[assertion](https://github.com/apache/incubator-mxnet/blob/master/python/mxnet/gluon/nn/conv_layers.py#L719)
 to check the layout format, the conv layers didn't check it. So if I I used a 
'NWC' layout,  the output error message is even more strange.  A possible 
solution is to add assertion into each conv1d, conv2d stuffs, and fix the 
corresponding documents.
   @zhreshold @mli Any suggestion for this? 
   
   Demo code:
   ```python
   from mxnet import nd
   from mxnet.gluon import nn
   
   x = nd.ones((5,2,3))
   conv = nn.Conv1D(layout='NWC', channels=4, kernel_size=3)
   conv.initialize()
   print(x.shape)
   print(conv(x).shape)
   ```
   Error Log:
   ```
   mxnet.base.MXNetError: Invalid Input: 'NWC', valid values are: {None, 
'NCDHW', 'NCHW', 'NCW', 'NDHWC', 'NHWC'}, in operator Convolution(name="", 
layout="NWC", pad="(0,)", num_filter="4", dilate="(1,)", num_group="1", 
stride="(1,)", no_bias="False", kernel="(3,)")
   ```


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 closed pull request #8582: Yolo2 operator

[GitBox]

zhreshold closed pull request #8582: Yolo2 operator
URL: https://github.com/apache/incubator-mxnet/pull/8582
 
 
   

This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:

As this is a foreign pull request (from a fork), the diff is supplied
below (as it won't show otherwise due to GitHub magic):

diff --git a/src/operator/contrib/yolo_output-inl.h 
b/src/operator/contrib/yolo_output-inl.h
new file mode 100644
index 000..9f95bcacdca
--- /dev/null
+++ b/src/operator/contrib/yolo_output-inl.h
@@ -0,0 +1,735 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+/*!
+ * \file yolo_output-inl.h
+ * \brief yolo-v2 output layer, v1 to be added
+ * \author Joshua Zhang
+*/
+#ifndef MXNET_OPERATOR_CONTRIB_YOLO_OUTPUT_INL_H_
+#define MXNET_OPERATOR_CONTRIB_YOLO_OUTPUT_INL_H_
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include "../operator_common.h"
+#include "../mshadow_op.h"
+#include "../mxnet_op.h"
+#include "../tensor/sort_op.h"
+
+namespace mxnet {
+namespace op {
+namespace yolo2out_enum {
+enum Yolo2OutputOpInputs {kData, kLabel};
+enum Yolo2OutputOpOutputs {kOut, kTemp, kCopy};
+enum Yolo2OutputOpAuxiliary {kCounter};
+enum Yolo2OutputOpResource {kTempSpace};
+}  // namespace yolo2out_enum
+
+struct Yolo2OutputParam : public dmlc::Parameter {
+  int num_class;
+  int num_anchor;
+  float overlap_thresh;
+  float object_grad_scale;
+  float background_grad_scale;
+  float class_grad_scale;
+  float coord_grad_scale;
+  nnvm::Tuple anchors;
+  int warmup_samples;
+  float warmup_grad_scale;
+  float nms_threshold;
+  int nms_topk;
+  bool force_suppress;
+  DMLC_DECLARE_PARAMETER(Yolo2OutputParam) {
+DMLC_DECLARE_FIELD(num_class).set_lower_bound(1)
+.describe("Number of object classes.");
+DMLC_DECLARE_FIELD(num_anchor).set_default(5)
+.set_lower_bound(1)
+.describe("Number of anchors.");
+DMLC_DECLARE_FIELD(overlap_thresh).set_default(0.6)
+.describe("Positive overlap threshold.");
+DMLC_DECLARE_FIELD(object_grad_scale).set_default(1.0)
+.describe("Gradient scale for positive objects.");
+DMLC_DECLARE_FIELD(background_grad_scale).set_default(1.0)
+.describe("Gradient scale for background.");
+DMLC_DECLARE_FIELD(class_grad_scale).set_default(1.0)
+.describe("Gradient scale for positive objects.");
+DMLC_DECLARE_FIELD(coord_grad_scale).set_default(1.0)
+.describe("Gradient scale for box offsets.");
+DMLC_DECLARE_FIELD(anchors)
+.set_default({1.08f, 1.19f, 3.42f, 4.41f, 6.63f, 11.38f, 9.42f, 5.11f, 
16.62f, 10.52f})
+.describe("Predefined anchor box widths and heights.");
+DMLC_DECLARE_FIELD(warmup_samples).set_default(12800)
+.describe("Number of images to warm up towards averaging position for box "
+"predictions when starting a new training. ");
+DMLC_DECLARE_FIELD(warmup_grad_scale).set_default(0.01)
+.describe("Gradient scale for non-critical anchors during warm-up stage.");
+DMLC_DECLARE_FIELD(nms_threshold).set_default(0.5f)
+.describe("Non-maximum suppression threshold.");
+DMLC_DECLARE_FIELD(force_suppress).set_default(false)
+.describe("Suppress all detections regardless of class_id.");
+DMLC_DECLARE_FIELD(nms_topk).set_default(-1)
+.describe("Keep maximum top k detections before nms, -1 for no limit.");
+  }
+};  // struct Yolo2OutputParam
+
+/*!
+   * \brief 1-D Intersection between two sections, either along x or y 
direction
+   *
+   * \param l1 left/top coordinate of box 1
+   * \param r1 right/bottom coordinate of box 1
+   * \param l2 left/top coordinate of box 2
+   * \param r2 right/bottom coordinate of box 2
+   * \tparam DType the data type
+   */
+template
+MSHADOW_XINLINE DType Intersect(DType l1, DType r1, DType l2, DType r2) {
+  DType left = l1 > l2 ? l1 : l2;
+  DType right = r1 < r2 ? r1 : r2;
+  DType w = right - left;
+  return w > 0 ? w : DType(0);
+}
+
+/*!
+   * \brief Box intersection-over-union ratio calculation
+   *
+   * \param l1 left coordinate of box
+   * \param t1 top coordinate of box
+   * \param r1 right coordinate of box
+   * 

[GitHub] zhreshold commented on issue #8582: Yolo2 operator

[GitBox]

zhreshold commented on issue #8582: Yolo2 operator
URL: https://github.com/apache/incubator-mxnet/pull/8582#issuecomment-386196536
 
 
   Closing this for now. Trying to implement in a new approach.


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 issue #8547: Revised validating process of a bounding box in ImageDetIter

[GitBox]

zhreshold commented on issue #8547: Revised validating process of a bounding 
box in ImageDetIter
URL: https://github.com/apache/incubator-mxnet/pull/8547#issuecomment-386196504
 
 
   Closing this for now. Let me know to reopen it if you got new ideas.


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 #10736: [MXNET-367] update mkldnn to v0.14 and disable building test examples

[GitBox]

szha commented on a change in pull request #10736: [MXNET-367] update mkldnn to 
v0.14 and disable building test examples
URL: https://github.com/apache/incubator-mxnet/pull/10736#discussion_r185700483
 
 

 ##
 File path: prepare_mkl.sh
 ##
 @@ -58,16 +58,16 @@ MXNET_ROOT=`dirname $0`
 USE_MKLML=0
 # NOTE: if you update the following line, please also update the dockerfile at
 # tests/ci_build/Dockerfile.mkl
-VERSION_MATCH=20171227
+VERSION_MATCH=20180406
 
 Review comment:
   I haven't seen such use before. Has this been added as an option to USE_BLAS 
flag?


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] ashokei commented on a change in pull request #10736: [MXNET-367] update mkldnn to v0.14 and disable building test examples

[GitBox]

ashokei commented on a change in pull request #10736: [MXNET-367] update mkldnn 
to v0.14 and disable building test examples
URL: https://github.com/apache/incubator-mxnet/pull/10736#discussion_r185699357
 
 

 ##
 File path: prepare_mkl.sh
 ##
 @@ -58,16 +58,16 @@ MXNET_ROOT=`dirname $0`
 USE_MKLML=0
 # NOTE: if you update the following line, please also update the dockerfile at
 # tests/ci_build/Dockerfile.mkl
-VERSION_MATCH=20171227
+VERSION_MATCH=20180406
 
 Review comment:
   this script can be used independent of mkldnn (for eg: mshadow with mklml 
blas api)


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] xinyu-intel commented on issue #10629: [MXNET-343]fix Mkldnn with msvc

[GitBox]

xinyu-intel commented on issue #10629: [MXNET-343]fix Mkldnn with msvc
URL: https://github.com/apache/incubator-mxnet/pull/10629#issuecomment-386194605
 
 
   @yajiedesign seems some conflicts occured.


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] xu2011 opened a new issue #10793: dataloader load custom dataset error

[GitBox]

xu2011 opened a new issue #10793: dataloader load custom dataset error
URL: https://github.com/apache/incubator-mxnet/issues/10793
 
 
   Python 3.6 on MacOS
   
   I try to load custom dataset with rec file using dataloader as following:
   `data = gluon.data.RecordFileDataset("face/test.rec")`
   `data_loader = gluon.data.DataLoader(data, batch_size=10, shuffle=True)`
   `for data, label in data_loader:`
   ` do(data)`
   
   ---
   KeyError  Traceback (most recent call last)
in ()
 1 data = gluon.data.RecordFileDataset("landmark/test.rec")
 2 data_loader = gluon.data.DataLoader(data, batch_size=10, 
shuffle=True)
   > 3 for data, label in data_loader:
 4 print(label)
 5 
   
   ~/anaconda3/lib/python3.6/site-packages/mxnet/gluon/data/dataloader.py in 
__iter__(self)
   200 if self._num_workers == 0:
   201 for batch in self._batch_sampler:
   --> 202 yield self._batchify_fn([self._dataset[idx] for idx 
in batch])
   203 return
   204 
   
   ~/anaconda3/lib/python3.6/site-packages/mxnet/gluon/data/dataloader.py in 
default_batchify_fn(data)
93 else:
94 data = np.asarray(data)
   ---> 95 return nd.array(data, dtype=data.dtype)
96 
97 
   
   ~/anaconda3/lib/python3.6/site-packages/mxnet/ndarray/utils.py in 
array(source_array, ctx, dtype)
   144 return _sparse_array(source_array, ctx=ctx, dtype=dtype)
   145 else:
   --> 146 return _array(source_array, ctx=ctx, dtype=dtype)
   147 
   148 
   
   ~/anaconda3/lib/python3.6/site-packages/mxnet/ndarray/ndarray.py in 
array(source_array, ctx, dtype)
  2242 except:
  2243 raise TypeError('source_array must be array like 
object')
   -> 2244 arr = empty(source_array.shape, ctx, dtype)
  2245 arr[:] = source_array
  2246 return arr
   
   ~/anaconda3/lib/python3.6/site-packages/mxnet/ndarray/ndarray.py in 
empty(shape, ctx, dtype)
  3413 if dtype is None:
  3414 dtype = mx_real_t
   -> 3415 return NDArray(handle=_new_alloc_handle(shape, ctx, False, 
dtype))
   
   ~/anaconda3/lib/python3.6/site-packages/mxnet/ndarray/ndarray.py in 
_new_alloc_handle(shape, ctx, delay_alloc, dtype)
   136 ctypes.c_int(ctx.device_id),
   137 ctypes.c_int(int(delay_alloc)),
   --> 138 ctypes.c_int(int(_DTYPE_NP_TO_MX[np.dtype(dtype).type])),
   139 ctypes.byref(hdl)))
   140 return hdl
   
   KeyError: 
   
   ** also, the original dataset don't have labels, I only want to get features 
with pre-trained models, does anyone know what is right way to do so?


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 #10736: [MXNET-367] update mkldnn to v0.14 and disable building test examples

[GitBox]

szha commented on a change in pull request #10736: [MXNET-367] update mkldnn to 
v0.14 and disable building test examples
URL: https://github.com/apache/incubator-mxnet/pull/10736#discussion_r185699109
 
 

 ##
 File path: ci/docker/install/ubuntu_mklml.sh
 ##
 @@ -21,5 +21,5 @@
 # the whole docker cache for the image
 
 set -ex
-wget --no-check-certificate -O /tmp/mklml.tgz 
https://github.com/intel/mkl-dnn/releases/download/v0.12/mklml_lnx_2018.0.1.20171227.tgz
-tar -zxvf /tmp/mklml.tgz && cp -rf mklml_*/* /usr/local/ && rm -rf mklml_*
\ No newline at end of file
+wget --no-check-certificate -O /tmp/mklml.tgz 
https://github.com/intel/mkl-dnn/releases/download/v0.14/mklml_lnx_2018.0.3.20180406.tgz
 
 Review comment:
   any way to automatically get this from 3rdparty/mkldnn? and this should 
really be managed by mkldnn because mxnet shouldn't need to directly depend on 
mklml.


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 #10736: [MXNET-367] update mkldnn to v0.14 and disable building test examples

[GitBox]

szha commented on a change in pull request #10736: [MXNET-367] update mkldnn to 
v0.14 and disable building test examples
URL: https://github.com/apache/incubator-mxnet/pull/10736#discussion_r185698982
 
 

 ##
 File path: prepare_mkl.sh
 ##
 @@ -58,16 +58,16 @@ MXNET_ROOT=`dirname $0`
 USE_MKLML=0
 # NOTE: if you update the following line, please also update the dockerfile at
 # tests/ci_build/Dockerfile.mkl
-VERSION_MATCH=20171227
+VERSION_MATCH=20180406
 
 Review comment:
   Since this version is tied to the specific mkldnn commit, should this script 
live in mkldnn instead?


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] ashokei commented on issue #10773: specify common arch for mkldnn build, so it can be reused on all arch

[GitBox]

ashokei commented on issue #10773: specify common arch for mkldnn build, so it 
can be reused on all arch
URL: https://github.com/apache/incubator-mxnet/pull/10773#issuecomment-386193952
 
 
   This PR can work for runtime switching. The tradeoff is potential 
performance loss (which i do not see in my testing, as mkldnn already does 
runtime switching/JIT for its performance critical compute kernels). 
   
   I am following up with mkl-dnn team,  currently they did not give us any 
exact measurement on performance impact , if any. 
   
   If we truly need runtime switching for mkl-dnn+mxnet, they  (mkldnn team) 
suggest we set `-march` to common baseline architecture.
   
   Another extra option we could explore, is to add this PR with a optional 
build flag to build for native arch if users know they do not need runtime 
switching (via an environment export variable as i mention above).
   


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] jeremiedb commented on issue #9602: MXNET not working in R

[GitBox]

jeremiedb commented on issue #9602: MXNET not working in R
URL: 
https://github.com/apache/incubator-mxnet/issues/9602#issuecomment-386192860
 
 
   Hi, the official repo hasn't been updated. In the meantime, you can install 
this version (CPU only for now): 
   ```
   
install.packages("https://github.com/jeremiedb/mxnet_winbin/raw/master/mxnet.zip;,
 repos = NULL)
   ```
   


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 #10629: [MXNET-343]fix Mkldnn with msvc

[GitBox]

szha commented on issue #10629: [MXNET-343]fix Mkldnn with msvc
URL: https://github.com/apache/incubator-mxnet/pull/10629#issuecomment-386192736
 
 
   Ping


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] ashokei commented on issue #10773: specify common arch for mkldnn build, so it can be reused on all arch

[GitBox]

ashokei commented on issue #10773: specify common arch for mkldnn build, so it 
can be reused on all arch
URL: https://github.com/apache/incubator-mxnet/pull/10773#issuecomment-386191597
 
 
   we can explore other options.  One solution is to do this if some 
environment variable is exported, maybe. @szha is that ok? we could use it only 
for CI , or other use-cases where users may build on one system and copy to 
different arch system.
   
   
   


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] zheng-da commented on issue #10104: [WIP][MXNET-107] Fused RNN implementation for CPU

[GitBox]

zheng-da commented on issue #10104: [WIP][MXNET-107] Fused RNN implementation 
for CPU
URL: https://github.com/apache/incubator-mxnet/pull/10104#issuecomment-386191155
 
 
   What is the status of this PR?


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 closed pull request #10790: Fix learning rate of ctc example

[GitBox]

szha closed pull request #10790: Fix learning rate of ctc example
URL: https://github.com/apache/incubator-mxnet/pull/10790
 
 
   

This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:

As this is a foreign pull request (from a fork), the diff is supplied
below (as it won't show otherwise due to GitHub magic):

diff --git a/example/ctc/hyperparams.py b/example/ctc/hyperparams.py
index 7289d19c03f..cdcb874b3a6 100644
--- a/example/ctc/hyperparams.py
+++ b/example/ctc/hyperparams.py
@@ -29,7 +29,7 @@ def __init__(self):
 self._eval_epoch_size = 3000
 self._batch_size = 128
 self._num_epoch = 100
-self._learning_rate = 0.001
+self._learning_rate = 0.01
 self._momentum = 0.9
 self._num_label = 4
 # Network hyper parameters


 


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] apshreyans opened a new issue #10791: Unable to install mxnet in R 3.5.0

[GitBox]

apshreyans opened a new issue #10791: Unable to install mxnet in R 3.5.0
URL: https://github.com/apache/incubator-mxnet/issues/10791
 
 
   Note: Providing complete information in the most concise form is the best 
way to get help. This issue template serves as the checklist for essential 
information to most of the technical issues and bug reports. For non-technical 
issues and feature requests, feel free to present the information in what you 
believe is the best form.
   
   For Q & A and discussion, please start a discussion thread at 
https://discuss.mxnet.io 
   
   ## Description
   Installation error on RStudio R 3.5.0
   Error:   unable to access index for repository 
https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/bin/windows/contrib/3.5:
 cannot open URL 
'https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/bin/windows/contrib/3.5/PACKAGES'
   
   **Command Run:**
   cran <- getOption("repos")
   cran["dmlc"] <- 
"https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/;
   options(repos = cran)
   install.packages("mxnet")
   
   
   (Brief description of the problem in no more than 2 sentences.)
   
   ## Environment info (Required)
   R version 3.5.0 (2018-04-23) -- "Joy in Playing"
   Copyright (C) 2018 The R Foundation for Statistical Computing
   Platform: x86_64-w64-mingw32/x64 (64-bit)
   
   Package used (Python/R/Scala/Julia):
   **(I'm using R 3.5.0)**
   
   For R user, please provide R `sessionInfo()`:
   **SessionInfo():**
   > sessionInfo()
   R version 3.5.0 (2018-04-23)
   Platform: x86_64-w64-mingw32/x64 (64-bit)
   Running under: Windows 7 x64 (build 7601) Service Pack 1
   
   Matrix products: default
   
   locale:
   [1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United 
States.1252   
   [3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C  

   [5] LC_TIME=English_United States.1252
   
   attached base packages:
   [1] stats graphics  grDevices utils datasets  methods   base 
   
   loaded via a namespace (and not attached):
   [1] compiler_3.5.0 tools_3.5.0  
   
   ## Build info (Required if built from source)
   
   Compiler (gcc/clang/mingw/visual studio):
   
   MXNet commit hash:
   (Paste the output of `git rev-parse HEAD` here.)
   
   Build config:
   (Paste the content of config.mk, or the build command.)
   
   ## Error Message:
   (Paste the complete error message, including stack trace.)
   
no package 'knitr' was found
   > install.packages('mxnet')
   Installing package into ‘\\homedrive$/Shreyans.Bathiyan/R/win-library/3.5’
   (as ‘lib’ is unspecified)
   Warning in install.packages :
 package ‘mxnet’ is not available (for R version 3.5.0)
   Error: invalid version specification ‘NA’
   In addition: Warning message:
   In utils:::packageDescription(packageName, fields = "Version") :
 no package 'knitr' was found
   > cran <- getOption("repos")
   > cran["dmlc"] <- 
"https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU;
   > options(repos = cran)
   > install.packages("mxnet")
   Error in install.packages : Updating loaded packages
   
   Restarting R session...
   
   > install.packages("mxnet")
   Installing package into ‘\\homedrive$/Shreyans.Bathiyan/R/win-library/3.5’
   (as ‘lib’ is unspecified)
   also installing the dependencies ‘bindr’, ‘brew’, ‘colorspace’, ‘utf8’, 
‘gtable’, ‘reshape2’, ‘lazyeval’, ‘assertthat’, ‘bindrcpp’, ‘pkgconfig’, ‘R6’, 
‘BH’, ‘plogr’, ‘digest’, ‘yaml’, ‘hms’, ‘XML’, ‘Rook’, ‘dichromat’, ‘plyr’, 
‘munsell’, ‘labeling’, ‘viridisLite’, ‘cli’, ‘crayon’, ‘pillar’, ‘tidyselect’, 
‘ggplot2’, ‘gridExtra’, ‘dplyr’, ‘downloader’, ‘glue’, ‘htmltools’, 
‘htmlwidgets’, ‘igraph’, ‘influenceR’, ‘purrr’, ‘RColorBrewer’, ‘readr’, 
‘rlang’, ‘rstudioapi’, ‘rgexf’, ‘scales’, ‘tibble’, ‘tidyr’, ‘viridis’, 
‘stringi’, ‘Rcpp’, ‘DiagrammeR’, ‘visNetwork’, ‘data.table’, ‘jsonlite’, 
‘magrittr’, ‘stringr’
   
   Warning in install.packages :
 unable to access index for repository 
https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/bin/windows/contrib/3.5:
 cannot open URL 
'https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/bin/windows/contrib/3.5/PACKAGES'
   
 There is a binary version available but the source version is later:
   binary source needs_compilation
   stringi  1.1.7  1.2.2  TRUE
   
 Binaries will be installed
   Package which is only available in source form, and may need compilation of
 C/C++/Fortran: ‘mxnet’
   
   
   Kindly help me installing mxnet package.


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 #10767: add reverse option to ndarray inplace reshape

[GitBox]

piiswrong commented on a change in pull request #10767: add reverse option to 
ndarray inplace reshape
URL: https://github.com/apache/incubator-mxnet/pull/10767#discussion_r185695101
 
 

 ##
 File path: python/mxnet/ndarray/ndarray.py
 ##
 @@ -1032,15 +1045,14 @@ def reshape(self, *shape, **kwargs):
 if len(kwargs) != 1:
 raise TypeError("Only 'shape' is supported as keyword 
argument. Got: {}."
 .format(', '.join(kwargs.keys(
-else:
-assert not kwargs,\
-"Specifying both positional and keyword arguments is not 
allowed in reshape."
+reverse = kwargs.get('reverse', False)
 
 Review comment:
   needs to make sure kwargs doesn't contain other arguments


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] jinhuang415 commented on issue #10783: Fix 'make clean USE_MKLDNN=1' will build mkldnn issue

[GitBox]

jinhuang415 commented on issue #10783: Fix 'make clean USE_MKLDNN=1' will build 
mkldnn issue
URL: https://github.com/apache/incubator-mxnet/pull/10783#issuecomment-386188207
 
 
   @szha @zheng-da I agree define a new build target for mkldnn and integrate 
into build dependency will be cleaner solution, I will investigate how to do 
that.


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] apshreyans commented on issue #9602: MXNET not working in R

[GitBox]

apshreyans commented on issue #9602: MXNET not working in R
URL: 
https://github.com/apache/incubator-mxnet/issues/9602#issuecomment-386187824
 
 
   Hi 
   I am unable to install the 'mxnet package' in R 3.5.0.
   Error below:
   
   **Start***
   Warning in install.packages :
 unable to access index for repository 
https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/bin/windows/contrib/3.5:
 cannot open URL 
'https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/R/CRAN/GPU/bin/windows/contrib/3.5/PACKAGES'
   
 There is a binary version available but the source version is later:
   binary source needs_compilation
   stringi  1.1.7  1.2.2  TRUE
   
 Binaries will be installed
   Package which is only available in source form, and may need compilation of
 C/C++/Fortran: ‘mxnet’
   **End
   
   Kindly help.
   Thanks much in advance.
   Shreyans


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 closed pull request #9142: add lstm sentimental analysis tutorial

[GitBox]

szha closed pull request #9142: add lstm sentimental analysis tutorial
URL: https://github.com/apache/incubator-mxnet/pull/9142
 
 
   

This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:

As this is a foreign pull request (from a fork), the diff is supplied
below (as it won't show otherwise due to GitHub magic):

diff --git a/example/lstm-sentimal_analysis/lstm_cell.png 
b/example/lstm-sentimal_analysis/lstm_cell.png
new file mode 100644
index 000..442b50b6f0f
Binary files /dev/null and b/example/lstm-sentimal_analysis/lstm_cell.png differ
diff --git a/example/lstm-sentimal_analysis/lstm_diagram.png 
b/example/lstm-sentimal_analysis/lstm_diagram.png
new file mode 100644
index 000..b2cebc14e0b
Binary files /dev/null and b/example/lstm-sentimal_analysis/lstm_diagram.png 
differ
diff --git a/example/lstm-sentimal_analysis/tutorial_lstm.ipynb 
b/example/lstm-sentimal_analysis/tutorial_lstm.ipynb
new file mode 100644
index 000..571288096b9
--- /dev/null
+++ b/example/lstm-sentimal_analysis/tutorial_lstm.ipynb
@@ -0,0 +1,491 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+"# LSTM: Sentimental Analysis using IMDB movie review data set\n",
+"\n",
+"- In this tutorial, we will create a model that determines whether movie 
reviews are positive or negative.\n",
+"- This tutorial is intended for those who have a good understanding of 
python and a brief understanding of the LSTM model.\n",
+"\n",
+"\n",
+"## Prerequisites\n",
+"\n",
+"You need to install \n",
+"- Python 3.6. https://www.python.org/downloads/\n;,
+"- mxnet(gpu ver.) 
https://mxnet.incubator.apache.org/get_started/install.html\n;,
+"\n",
+"You also need to understand the LSTM model.  \n",
+"In this tutorial, a simple knowledge already presented on several 
websites is sufficient. However, if you want a deep understanding, please refer 
to the following article.\n",
+"- https://arxiv.org/pdf/1503.00185v5.pdf\n;,
+"\n",
+"Finally, you need to know about python, mxnet, gloun package, and so 
on.\n",
+"\n",
+"\n",
+"## The Data\n",
+"\n",
+"We will use IMDB. You can download it from the following link.  \n",
+"It is divided into 25000 train data and test data each labeled.  \n",
+"See the data set's README for more information.\n",
+"- http://ai.stanford.edu/~amaas/data/sentiment/\n;,
+"\n",
+"We will use Stanford's Global Vector for Word Representation (GloVe) 
embedding for word embedding.  \n",
+"Download glove.42B.300d.zip from the following link:\n",
+"- https://nlp.stanford.edu/projects/glove/\n;,
+"\n",
+"\n",
+"## Prepare the Data\n",
+"\n",
+"# 1. Read the data and label it. (1 means positive and 0 means 
negative.)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+"import os\n",
+"\n",
+"# helper function to read files\n",
+"def read_files(folder_name):\n",
+"sentiments =[]\n",
+"filenames = os.listdir(os.curdir+\"/\"+folder_name)\n",
+"\n",
+"for file in filenames:\n",
+"with open(folder_name+\"/\"+file, \"r\", encoding=\"utf8\") as 
f:\n",
+"data = f.read().replace(\"\\n\", \"\")\n",
+"sentiments.append(data)\n",
+"\n",
+"return sentiments\n",
+"\n",
+"data_path = '../data/aclImdb'\n",
+"\n",
+"# prepare train data\n",
+"train_pos_foldername = data_path + \"/train/pos\"\n",
+"train_pos_sentiments = read_files(train_pos_foldername)\n",
+"\n",
+"train_neg_foldername = data_path + \"/train/neg\"\n",
+"train_neg_sentiments = read_files(train_neg_foldername)\n",
+"\n",
+"train_pos_labels = [1 for _ in train_pos_sentiments]\n",
+"train_neg_labels = [0 for _ in train_neg_sentiments]\n",
+"train_all_labels = train_pos_labels + train_neg_labels\n",
+"\n",
+"train_all_sentiments = train_pos_sentiments + train_neg_sentiments\n",
+"\n",
+"# prepare test data\n",
+"test_pos_foldername = data_path + \"/test/pos\"\n",
+"test_pos_sentiments = read_files(test_pos_foldername)\n",
+"\n",
+"test_neg_foldername = data_path + \"/test/neg\"\n",
+"test_neg_sentiments = read_files(test_neg_foldername)\n",
+"\n",
+"test_pos_labels = [1 for _ in test_pos_sentiments]\n",
+"test_neg_labels = [0 for _ in test_neg_sentiments]\n",
+"test_all_labels = test_pos_labels + test_neg_labels\n",
+"\n",
+"test_all_sentiments = train_pos_sentiments + train_neg_sentiments"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+"# 2. We will make a dictionary from the words in the train data. 
Before that, let's define some helper functions."
+  

[GitHub] szha commented on issue #9142: add lstm sentimental analysis tutorial

[GitBox]

szha commented on issue #9142: add lstm sentimental analysis tutorial
URL: https://github.com/apache/incubator-mxnet/pull/9142#issuecomment-386187469
 
 
   Closing for lack of activity. Please don't hesitate to reopen one if you 
think all concerns such as place for contribution is addressed.


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 #10354: Expose the number of GPUs.

[GitBox]

szha commented on issue #10354: Expose the number of GPUs.
URL: https://github.com/apache/incubator-mxnet/pull/10354#issuecomment-386187188
 
 
   Ping


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] zheng-da commented on issue #10783: Fix 'make clean USE_MKLDNN=1' will build mkldnn issue

[GitBox]

zheng-da commented on issue #10783: Fix 'make clean USE_MKLDNN=1' will build 
mkldnn issue
URL: https://github.com/apache/incubator-mxnet/pull/10783#issuecomment-386186212
 
 
   Just tried it again. It doesn't work if someone wants to build "make test".
   


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] ashokei commented on issue #10773: specify common arch for mkldnn build, so it can be reused on all arch

[GitBox]

ashokei commented on issue #10773: specify common arch for mkldnn build, so it 
can be reused on all arch
URL: https://github.com/apache/incubator-mxnet/pull/10773#issuecomment-386185282
 
 
   @TaoLv @szha I measured the performance with this PR change on inference 
models using benchmark_score script. I did not see any performance difference. 
   
   Intel mkl-dnn team recommended this approach when i asked them about 
"similar option to DYNAMIC_ARCH in openblas" for mkldnn.  
   
   mklml library (which is shipped as part of mkldnn) uses this approach.
   
   @zheng-da This flag will ensure a common baseline arch is used, instead of 
`native` or default arch of the system used for building mkldnn. This will just 
ensure the auxiliary/support code does not cause illegal instruction/crash etc. 
The actual mkldnn kernels/primitives are generated at runtime based on 
underlying arch of the system being used for running.


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] zheng-da commented on a change in pull request #10783: Fix 'make clean USE_MKLDNN=1' will build mkldnn issue

[GitBox]

zheng-da commented on a change in pull request #10783: Fix 'make clean 
USE_MKLDNN=1' will build mkldnn issue
URL: https://github.com/apache/incubator-mxnet/pull/10783#discussion_r185692668
 
 

 ##
 File path: Makefile
 ##
 @@ -66,13 +66,22 @@ $(warning "USE_MKL2017 is deprecated. We will switch to 
USE_MKLDNN.")
 endif
 
 ifeq ($(USE_MKLDNN), 1)
+ifeq ($(MAKECMDGOALS),)
 
 Review comment:
   This is really a cool solution. Thanks. Learn some new stuff.


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 #10773: specify common arch for mkldnn build, so it can be reused on all arch

[GitBox]

szha commented on issue #10773: specify common arch for mkldnn build, so it can 
be reused on all arch
URL: https://github.com/apache/incubator-mxnet/pull/10773#issuecomment-386184711
 
 
   @TaoLv that means we need to at least compromise the performance of the 
pre-built binary, and I'd rather we don't go down that route. can we instead 
make a similar option to DYNAMIC_ARCH in openblas?


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] chinakook opened a new pull request #10790: Fix learning rate of ctc example

[GitBox]

chinakook opened a new pull request #10790: Fix learning rate of ctc example
URL: https://github.com/apache/incubator-mxnet/pull/10790
 
 
   ## Description ##
   The training of ctc will not converge in 100 epochs with learning rate 
0.001, So change it to 0.01.
   
   ## Checklist ##
   ### Essentials ###
   Please feel free to remove inapplicable items for your PR.
   - [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to 
the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) 
created (except PRs with tiny changes)
   - [ ] Changes are complete (i.e. I finished coding on this PR)
   - [ ] All changes have test coverage:
   - Unit tests are added for small changes to verify correctness (e.g. adding 
a new operator)
   - Nightly tests are added for complicated/long-running ones (e.g. changing 
distributed kvstore)
   - Build tests will be added for build configuration changes (e.g. adding a 
new build option with NCCL)
   - [ ] Code is well-documented: 
   - For user-facing API changes, API doc string has been updated. 
   - For new C++ functions in header files, their functionalities and arguments 
are documented. 
   - For new examples, README.md is added to explain the what the example does, 
the source of the dataset, expected performance on test set and reference to 
the original paper if applicable
   - Check the API doc at 
http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
   - [ ] To the my best knowledge, examples are either not affected by this 
change, or have been fixed to be compatible with this change
   
   ### Changes ###
   - [ ] Feature1, tests, (and when applicable, API doc)
   - [ ] Feature2, tests, (and when applicable, API doc)
   
   ## Comments ##
   - If this change is a backward incompatible change, why must this change be 
made.
   - Interesting edge cases to note here
   


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] anirudhacharya commented on issue #10789: Reshape of input array when the shape is available only at runtime is not possible

[GitBox]

anirudhacharya commented on issue #10789: Reshape of input array when the shape 
is available only at runtime is not possible 
URL: 
https://github.com/apache/incubator-mxnet/issues/10789#issuecomment-386179606
 
 
   Proposed Solution:
   
   A potential solution to this is to modify the existing ``reshape_like`` 
operator to accept another boolean parameter like ``infer_shape``. And based on 
this parameter we can either infer the shape of second input( as it is done 
now) or take the second input as the shape array itself.
   
   Caffe2 does something similar in their 
[resizeLike](https://caffe2.ai/docs/operators-catalogue.html#resizelike) 
operator. 
   
   @haojin2 @nswamy @piiswrong @zheng-da @eric-haibin-lin 


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] wangshuailong commented on issue #10294: [Help] How I can change the in_data in CustomOp

[GitBox]

wangshuailong commented on issue #10294: [Help] How I can change the in_data in 
CustomOp
URL: 
https://github.com/apache/incubator-mxnet/issues/10294#issuecomment-386177858
 
 
   @anirudhacharya Didn't try discuss.mxnet.io before, looks awesome! I will 
try! 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] wangshuailong closed issue #10294: [Help] How I can change the in_data in CustomOp

[GitBox]

wangshuailong closed issue #10294: [Help] How I can change the in_data in 
CustomOp
URL: https://github.com/apache/incubator-mxnet/issues/10294
 
 
   


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] jinhuang415 commented on a change in pull request #10783: Fix 'make clean USE_MKLDNN=1' will build mkldnn issue

[GitBox]

jinhuang415 commented on a change in pull request #10783: Fix 'make clean 
USE_MKLDNN=1' will build mkldnn issue
URL: https://github.com/apache/incubator-mxnet/pull/10783#discussion_r185687263
 
 

 ##
 File path: Makefile
 ##
 @@ -66,13 +66,22 @@ $(warning "USE_MKL2017 is deprecated. We will switch to 
USE_MKLDNN.")
 endif
 
 ifeq ($(USE_MKLDNN), 1)
+ifeq ($(MAKECMDGOALS),)
 
 Review comment:
   MAKECMDGOALS is a makefile builtin variable used to indicate the make 
target, if user input "make clean", MAKECMDGOALS will be 'clean', if user input 
"make all", MAKECMDGOALS  will be 'all', and default will be empty (we usually 
don't specify the target when make MXNet so MAKECMDGOALS will be empty and 
Makefile will find first target which is all to build).
   So user don't need to set MAKECMDGOALS, it is set automatically by make.


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] pribadihcr commented on issue #10775: RGB to gray ops

[GitBox]

pribadihcr commented on issue #10775: RGB to gray ops
URL: 
https://github.com/apache/incubator-mxnet/issues/10775#issuecomment-386174452
 
 
   @roywei,
   
   e.g the shape of tensor_image_rgb  is [1,3, 100, 100]
   rgb_weights = mx.nd.array([0.2989, 0.5870, 0.1140]).as_in_context(ctx) # the 
shape is [3, 1]
   tensor_image_rgb = mx.nd.transpose(tensor_image_rgb, (0, 2, 3, 1))  # make 
shape become [1, 100, 100, 3]
   tensor_image_gray = mx.nd.dot(tensor_image_rgb, rgb_weights) # the shape 
become  [1, 100, 100, 1]
   tensor_image_gray = mx.nd.transpose(tensor_image_gray, (0, 3, 1, 2)) # the 
shape become  [1, 1, 100, 100]


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] chaoer commented on issue #10745: Build from sources with Cuda 9.0 cudnn 7.1 on Window 10.

[GitBox]

chaoer commented on issue #10745: Build from sources with Cuda 9.0 cudnn 7.1 on 
Window 10.
URL: 
https://github.com/apache/incubator-mxnet/issues/10745#issuecomment-386174051
 
 
   @larroy @nswamy 
   Finally I find an alternative way to skip this problem.
   
   Step1:  using CMake-gui do the normal configure process;
   Step2:  find the CUDA option, an select the "CUDA_ARCH_NAME" to "Manual"; 
then click configuration;
   Step3: change the new appear( usually in red in the GUI) to what you want. 
(for example: sm_70); then click configuration and generate;
   Step4: compile the generated sln file using VS IDE.


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] chaoer closed issue #10745: Build from sources with Cuda 9.0 cudnn 7.1 on Window 10.

[GitBox]

chaoer closed issue #10745: Build from sources with Cuda 9.0 cudnn 7.1 on 
Window 10.
URL: https://github.com/apache/incubator-mxnet/issues/10745
 
 
   


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] TaoLv commented on issue #10773: specify common arch for mkldnn build, so it can be reused on all arch

[GitBox]

TaoLv commented on issue #10773: specify common arch for mkldnn build, so it 
can be reused on all arch
URL: https://github.com/apache/incubator-mxnet/pull/10773#issuecomment-386173398
 
 
   I‘m afraid this will still imapct the performance of non-jit code if arch is 
hard coded here. Can we use a flag like `NATIVE` to control compilation during 
make? If the flag is True, we should compile mxnet/mkldnn according to the 
native machine. Otherwise, only a minimum instruction set should be supported. 
Anyway, the flag would be default to True.
   
   @szha Do we need to define a subset of instruction or arch which mxnet 
should support? I notice that there is also a discussion on dev@ about removing 
ancient gpu arch support from mxnet.


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-site] branch asf-site updated: Bump the publish timestamp.

[zhasheng]

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

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


The following commit(s) were added to refs/heads/asf-site by this push:
 new 94af6d0  Bump the publish timestamp.
94af6d0 is described below

commit 94af6d083db88c9401bfc37af36f7fedd31c6d06
Author: mxnet-ci 
AuthorDate: Thu May 3 01:56:21 2018 +

Bump the publish timestamp.
---
 date.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/date.txt b/date.txt
new file mode 100644
index 000..39a0a8b
--- /dev/null
+++ b/date.txt
@@ -0,0 +1 @@
+Thu May  3 01:56:21 UTC 2018

-- 
To stop receiving notification emails like this one, please contact
zhash...@apache.org.

[GitHub] roywei commented on issue #10775: RGB to gray ops

[GitBox]

roywei commented on issue #10775: RGB to gray ops
URL: 
https://github.com/apache/incubator-mxnet/issues/10775#issuecomment-386170200
 
 
   @pribadihcr , no we currently rely on open cv's implementation, you could 
follow tf.image.rgb_to_grayscale's 
[implementation](https://github.com/tensorflow/tensorflow/blob/r1.8/tensorflow/python/ops/image_ops_impl.py#L1278)
 to do something similar on NDArrays. Refer to 
[expand_dim](https://mxnet.incubator.apache.org/api/python/ndarray/ndarray.html#mxnet.ndarray.NDArray.expand_dims),
 and maybe convert to numpy array and use 
[tensordot](https://docs.scipy.org/doc/numpy-1.14.0/reference/generated/numpy.tensordot.html)
 if necessary. (not sure if mx.nd.dot is same as tensordot with axis =[-1, -1])
   
   @sandeep-krishnamurthy could you help label feature request, NDArray? 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] roywei commented on issue #10775: RGB to gray ops

[GitBox]

roywei commented on issue #10775: RGB to gray ops
URL: 
https://github.com/apache/incubator-mxnet/issues/10775#issuecomment-386170200
 
 
   @pribadihcr , no we currently rely on open cv's implementation, you could 
follow tf.image.rgb_to_grayscale's 
[implementation](https://github.com/tensorflow/tensorflow/blob/r1.8/tensorflow/python/ops/image_ops_impl.py#L1278)
 to do something similar on NDArrays. Refer to 
[expand_dim](https://mxnet.incubator.apache.org/api/python/ndarray/ndarray.html#mxnet.ndarray.NDArray.expand_dims),
 and maybe convert to numpy array and use 
[tensordot](https://docs.scipy.org/doc/numpy-1.14.0/reference/generated/numpy.tensordot.html)
 if necessary.
   
   @sandeep-krishnamurthy could you help label feature request, NDArray? 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] pribadihcr commented on issue #10775: RGB to gray ops

[GitBox]

pribadihcr commented on issue #10775: RGB to gray ops
URL: 
https://github.com/apache/incubator-mxnet/issues/10775#issuecomment-386168389
 
 
   @roywei,
   
   I think I know how to resolve, I was looking up the source code of  
tf.image.rgb_to_grayscale. They use dot operation with the weight rgb [0.2989, 
0.5870, 0.1140]. So, may be I will use mx.nd.dot. 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] pribadihcr closed issue #10775: RGB to gray ops

[GitBox]

pribadihcr closed issue #10775: RGB to gray ops
URL: https://github.com/apache/incubator-mxnet/issues/10775
 
 
   


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] huyangc opened a new issue #10788: Start a process for training. The training get stuck

[GitBox]

huyangc opened a new issue #10788: Start a process for training. The training 
get stuck
URL: https://github.com/apache/incubator-mxnet/issues/10788
 
 
   I start a process for training procedure, including loading data, prepare a 
module, and module.fit. However, when I just ran the training procedure, it can 
be successfully started, but when I start a process of the training procedure, 
it get stuck with the gpu memory allocated but GPU-Util is also 0%.
   
   some pseudocode here
   ```
   def train_net(args):
   sym = build_net()
   mod = mx.mod.Module(sym, ctx=gpu(0-7))
   initializer=mx.xiavier
   dataiter = get_dataiter()
   mod.fit(...)
   
   def train():
   p = mp.Process(target=train_net, args=(args,))
   p.start()
   p.join()
   
   ```
   


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] pribadihcr commented on issue #10775: RGB to gray ops

[GitBox]

pribadihcr commented on issue #10775: RGB to gray ops
URL: 
https://github.com/apache/incubator-mxnet/issues/10775#issuecomment-386165325
 
 
   @roywei,
   
   I mean the input from a tensor like tf.image.rgb_to_grayscale. the imdecode 
need the input  in int8 buffer. or do you have any example implementation?


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] pribadihcr commented on issue #10775: RGB to gray ops

[GitBox]

pribadihcr commented on issue #10775: RGB to gray ops
URL: 
https://github.com/apache/incubator-mxnet/issues/10775#issuecomment-386165325
 
 
   @roywei,
   
   I mean the input from a tensor like tf.image.rgb_to_grayscale. the imdecode 
hte input need in int8 buffer. or do you have any example implementation?


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] aaronmarkham commented on issue #10622: [MXNET-341] Added General Info to API Docs for Python and Gluon

[GitBox]

aaronmarkham commented on issue #10622: [MXNET-341] Added General Info to API 
Docs for Python and Gluon
URL: https://github.com/apache/incubator-mxnet/pull/10622#issuecomment-386164224
 
 
   GTG! Ship 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] leezu commented on issue #10506: A todo list for the sparse feature (GPU)

[GitBox]

leezu commented on issue #10506: A todo list for the sparse feature (GPU)
URL: 
https://github.com/apache/incubator-mxnet/issues/10506#issuecomment-386158560
 
 
   - [ ] Standalone pruning operator as in 
https://github.com/apache/incubator-mxnet/blob/274fae82d66b1cfed48f78e69538631c57c5b6b6/src/operator/tensor/cast_storage-inl.cuh#L116-L153


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] thomelane commented on issue #10622: [MXNET-341] Added General Info to API Docs for Python and Gluon

[GitBox]

thomelane commented on issue #10622: [MXNET-341] Added General Info to API Docs 
for Python and Gluon
URL: https://github.com/apache/incubator-mxnet/pull/10622#issuecomment-386156283
 
 
   @aaronmarkham what's the latest with changes of gluon homepage? good to 
merge 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] ankkhedia commented on issue #10171: CoreML tool, wrong mlmodelc file name for Inception-BN

[GitBox]

ankkhedia commented on issue #10171: CoreML tool, wrong mlmodelc file name for 
Inception-BN
URL: 
https://github.com/apache/incubator-mxnet/issues/10171#issuecomment-386154990
 
 
   @sandeep-krishnamurthy  Could you please label this issue as converter and 
bug


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] ankkhedia commented on issue #10217: Building with OpenCV causes link errors

[GitBox]

ankkhedia commented on issue #10217: Building with OpenCV causes link errors
URL: 
https://github.com/apache/incubator-mxnet/issues/10217#issuecomment-386154628
 
 
   @rahul003 Did the above suggestion worked for you? 
   @sandeep-krishnamurthy  Could you please tag the issue as Pending Requester 
Info and build


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] ankkhedia commented on issue #10187: OpenCV 3.3.1 undefined reference to compiler CXX_3.4.21

[GitBox]

ankkhedia commented on issue #10187: OpenCV 3.3.1 undefined reference to 
compiler CXX_3.4.21
URL: 
https://github.com/apache/incubator-mxnet/issues/10187#issuecomment-386153970
 
 
   @sandeep-krishnamurthy Could you please close this issue and tag the issue 
as question, build


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] ankkhedia commented on issue #10187: OpenCV 3.3.1 undefined reference to compiler CXX_3.4.21

[GitBox]

ankkhedia commented on issue #10187: OpenCV 3.3.1 undefined reference to 
compiler CXX_3.4.21
URL: 
https://github.com/apache/incubator-mxnet/issues/10187#issuecomment-386153970
 
 
   @sandeep-krishnamurthy Could you please close this issue and tag the issue 
as question and build


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] ankkhedia commented on issue #10174: Install v0.10.0 errors

[GitBox]

ankkhedia commented on issue #10174: Install v0.10.0 errors
URL: 
https://github.com/apache/incubator-mxnet/issues/10174#issuecomment-386153815
 
 
   @sandeep-krishnamurthy could you please label the issue Pending requester 
info


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] haojin2 commented on a change in pull request #10645: [MXNET-354] Support elemwise_add/sub between dense and row sparse tensors on CPU/GPU

[GitBox]

haojin2 commented on a change in pull request #10645: [MXNET-354] Support 
elemwise_add/sub between dense and row sparse tensors on CPU/GPU
URL: https://github.com/apache/incubator-mxnet/pull/10645#discussion_r185669065
 
 

 ##
 File path: src/operator/tensor/elemwise_binary_op-inl.h
 ##
 @@ -450,6 +493,91 @@ void ElemwiseBinaryOp::DnsCsrDnsOp(mshadow::Stream 
*s,
   });
 }
 
+/*!
+ * \brief Kernel for performing elemwise op between dense and rsp tensor
+ * \param iglobal thread id
+ * \param req  type of request
+ * \param out  output array
+ * \param dns_data data array of dense input
+ * \param rsp_data data array of rsp input
+ * \param rsp_indices  indices array of rsp input
+ * \param num_rows number of rows of both inputs
+ * \param nz_rows  number of non-zero rows of rsp tensor
+ * \param num_cols number of columns of both inputs
+ */
+template
+struct ElemwiseDnsRspDnsKernel {
+  template
+  MSHADOW_XINLINE static void Map(int i, DType* out, DType* dns_data,
+  const DType* rsp_data, const IType* 
rsp_indices,
+  const nnvm::dim_t num_rows, const 
nnvm::dim_t nz_rows,
+  const nnvm::dim_t num_cols) {
+if (i < nz_rows * num_cols) {
+  const nnvm::dim_t rsp_idx = i / num_cols;
+  const nnvm::dim_t dns_row = rsp_indices[rsp_idx];
+  const nnvm::dim_t col = i % num_cols;
+  KERNEL_ASSIGN(out[dns_row * num_cols + col], req,
+OP::Map(dns_data[dns_row * num_cols + col],
+rsp_data[rsp_idx * num_cols + col]));
+}
+  }
+};
+
+/*! \brief DNS -op- RSP binary operator for non-canonical NDArray */
+template
+void ElemwiseBinaryOp::DnsRspDnsOp(mshadow::Stream *s,
+   const nnvm::NodeAttrs ,
+   const OpContext ,
+   const NDArray ,
+   const NDArray ,
+   const OpReqType req,
+   const NDArray ,
+   const bool reverse) {
+  using namespace mshadow;
+  using namespace mxnet_op;
+  CHECK_EQ(dns.storage_type(), kDefaultStorage);
+  CHECK_EQ(rsp.storage_type(), kRowSparseStorage);
+  CHECK_EQ(output.data().Size(), dns.data().Size());
+  CHECK(req != kAddTo);
+  CHECK(req != kNullOp);
+  const bool supported_op = std::is_same::value ||
+std::is_same::value;
+  CHECK(supported_op == true);
+  const nnvm::dim_t num_rows = dns.shape()[0];
+  const nnvm::dim_t num_cols = dns.data().Size() / num_rows;
+  const nnvm::dim_t nz_rows = rsp.aux_shape(rowsparse::kIdx).Size();
 
 Review comment:
   Done.


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] ashokei commented on issue #10773: specify common arch for mkldnn build, so it can be reused on all arch

[GitBox]

ashokei commented on issue #10773: specify common arch for mkldnn build, so it 
can be reused on all arch
URL: https://github.com/apache/incubator-mxnet/pull/10773#issuecomment-386153872
 
 
   If we build on same machine, and use on same machine then there is no 
problem. I notice CI , or other use-cases where we may be building on one 
machine and copying mkldnn lib along with mxnet to other machine with differnt 
arch. This PR will make the build be compatible by using a common baseline.
   
   There will not be any performance impact, as the mkldnn kernels/primitives 
are generated runtime based on underlying arch. This PR is only to make the 
build itself compatible across systems.


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] haojin2 commented on a change in pull request #10645: [MXNET-354] Support elemwise_add/sub between dense and row sparse tensors on CPU/GPU

[GitBox]

haojin2 commented on a change in pull request #10645: [MXNET-354] Support 
elemwise_add/sub between dense and row sparse tensors on CPU/GPU
URL: https://github.com/apache/incubator-mxnet/pull/10645#discussion_r185668916
 
 

 ##
 File path: src/operator/tensor/elemwise_binary_op-inl.h
 ##
 @@ -430,17 +471,19 @@ void ElemwiseBinaryOp::DnsCsrDnsOp(mshadow::Stream 
*s,
   MSHADOW_IDX_TYPE_SWITCH(csr_indptr.type_flag_, CType, {
 
 Review comment:
   Done.


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] haojin2 commented on a change in pull request #10645: [MXNET-354] Support elemwise_add/sub between dense and row sparse tensors on CPU/GPU

[GitBox]

haojin2 commented on a change in pull request #10645: [MXNET-354] Support 
elemwise_add/sub between dense and row sparse tensors on CPU/GPU
URL: https://github.com/apache/incubator-mxnet/pull/10645#discussion_r185668098
 
 

 ##
 File path: src/operator/tensor/elemwise_binary_op-inl.h
 ##
 @@ -450,6 +493,91 @@ void ElemwiseBinaryOp::DnsCsrDnsOp(mshadow::Stream 
*s,
   });
 }
 
+/*!
+ * \brief Kernel for performing elemwise op between dense and rsp tensor
+ * \param iglobal thread id
+ * \param req  type of request
+ * \param out  output array
+ * \param dns_data data array of dense input
+ * \param rsp_data data array of rsp input
+ * \param rsp_indices  indices array of rsp input
+ * \param num_rows number of rows of both inputs
+ * \param nz_rows  number of non-zero rows of rsp tensor
+ * \param num_cols number of columns of both inputs
+ */
+template
+struct ElemwiseDnsRspDnsKernel {
+  template
+  MSHADOW_XINLINE static void Map(int i, DType* out, DType* dns_data,
+  const DType* rsp_data, const IType* 
rsp_indices,
+  const nnvm::dim_t num_rows, const 
nnvm::dim_t nz_rows,
+  const nnvm::dim_t num_cols) {
+if (i < nz_rows * num_cols) {
+  const nnvm::dim_t rsp_idx = i / num_cols;
+  const nnvm::dim_t dns_row = rsp_indices[rsp_idx];
+  const nnvm::dim_t col = i % num_cols;
+  KERNEL_ASSIGN(out[dns_row * num_cols + col], req,
+OP::Map(dns_data[dns_row * num_cols + col],
+rsp_data[rsp_idx * num_cols + col]));
+}
+  }
+};
+
+/*! \brief DNS -op- RSP binary operator for non-canonical NDArray */
+template
+void ElemwiseBinaryOp::DnsRspDnsOp(mshadow::Stream *s,
+   const nnvm::NodeAttrs ,
+   const OpContext ,
+   const NDArray ,
+   const NDArray ,
+   const OpReqType req,
+   const NDArray ,
+   const bool reverse) {
+  using namespace mshadow;
+  using namespace mxnet_op;
+  CHECK_EQ(dns.storage_type(), kDefaultStorage);
+  CHECK_EQ(rsp.storage_type(), kRowSparseStorage);
+  CHECK_EQ(output.data().Size(), dns.data().Size());
+  CHECK(req != kAddTo);
+  CHECK(req != kNullOp);
 
 Review comment:
   Done.


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] ankkhedia commented on issue #10167: HybridBlock can be slower than Block

[GitBox]

ankkhedia commented on issue #10167: HybridBlock can be slower than Block
URL: 
https://github.com/apache/incubator-mxnet/issues/10167#issuecomment-386152438
 
 
   @sandeep-krishnamurthy  Could you please label this issue as discussion


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] lanking520 commented on issue #10660: [MXNET-357] New Scala API Design (Symbol)

[GitBox]

lanking520 commented on issue #10660: [MXNET-357] New Scala API Design (Symbol)
URL: https://github.com/apache/incubator-mxnet/pull/10660#issuecomment-386150169
 
 
   Now is turn out to be fun. I removed the underscore filter and adding 
support for underscore function generation. Please kindly review the code 
parser section and find if there are possible ways to convert some "Any"s to 
actual Scala types


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] lanking520 opened a new pull request #10787: [MXNET-357][WIP] New Scala API Design (NDArray)

[GitBox]

lanking520 opened a new pull request #10787: [MXNET-357][WIP] New Scala API 
Design (NDArray)
URL: https://github.com/apache/incubator-mxnet/pull/10787
 
 
   ## Description ##
   See [full design 
document](https://cwiki.apache.org/confluence/display/MXNET/MXNet+Scala+API+Usability+Improvement)
   @nswamy @yzhliu 
   This PR is the Addition for new NDArray functions of Scala API
   Warning: current PR expose underscore functions
   
   ## Checklist ##
   ### Essentials ###
   Please feel free to remove inapplicable items for your PR.
   - [ ] Use QuasiQuote to replace original API Implementation (Reduce lines)
   - [ ] MakeAtomicFunction change to support String type parameter type in
   - [ ] User New namespace for New API (temporarily Symbol.api.function_name)
   - [ ] Default args using None to pass in as default
   - [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to 
the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) 
created (except PRs with tiny changes)
   - [ ] Changes are complete (i.e. I finished coding on this PR)
   - [ ] All changes have test coverage:
   - Unit tests are added for small changes to verify correctness (e.g. adding 
a new operator)
   - Nightly tests are added for complicated/long-running ones (e.g. changing 
distributed kvstore)
   - Build tests will be added for build configuration changes (e.g. adding a 
new build option with NCCL)
   - [ ] Code is well-documented: 
   - For user-facing API changes, API doc string has been updated. 
   - For new C++ functions in header files, their functionalities and arguments 
are documented. 
   - For new examples, README.md is added to explain the what the example does, 
the source of the dataset, expected performance on test set and reference to 
the original paper if applicable
   - Check the API doc at 
http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
   - [ ] To the my best knowledge, examples are either not affected by this 
change, or have been fixed to be compatible 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 issue #10767: add reverse option to ndarray inplace reshape

[GitBox]

szha commented on issue #10767: add reverse option to ndarray inplace reshape
URL: https://github.com/apache/incubator-mxnet/pull/10767#issuecomment-386142311
 
 
   ping @piiswrong 


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] anirudhacharya commented on issue #10316: MultiBoxDetection cannot pass consistency check

[GitBox]

anirudhacharya commented on issue #10316: MultiBoxDetection cannot pass 
consistency check
URL: 
https://github.com/apache/incubator-mxnet/issues/10316#issuecomment-386137115
 
 
   @nswamy @sandeep-krishnamurthy Please label this - "Operator", "Test"


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] leezu commented on issue #10768: Use numpy in RandomSampler

[GitBox]

leezu commented on issue #10768: Use numpy in RandomSampler
URL: https://github.com/apache/incubator-mxnet/pull/10768#issuecomment-386136121
 
 
   Thanks @asitstands . Comparing the numpy code to the mx.nd code you provided 
results in the following performance on my machine:
   ```
   
   In [3]: %timeit list(sample_mx(1529*8192))
   2.17 s ± 188 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
   
   In [4]: %timeit list(sample_np(1529*8192))
   1.3 s ± 73.2 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
   
   ```
   
   So relying on mx.nd.random.shuffle + asnumpy seems to add an extra second.
   
   Regarding RNG, our test cases set both numpy and mxnet seeds. I believe 
other parts of mxnet also use numpy random, so it may be good to document that 
both seeds must be set to get deterministic behavior. If this is the only place 
numpy.random is used it may be worth the extra second to stay consistent?


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] anirudhacharya commented on issue #10301: Disable async forward pass for fastest prediction?

[GitBox]

anirudhacharya commented on issue #10301: Disable async forward pass for 
fastest prediction?
URL: 
https://github.com/apache/incubator-mxnet/issues/10301#issuecomment-386136183
 
 
   @nswamy @sandeep-krishnamurthy @eric-haibin-lin Please label this - 
"Question" and close the issue, as it seems 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] anirudhacharya commented on issue #10294: [Help] How I can change the in_data in CustomOp

[GitBox]

anirudhacharya commented on issue #10294: [Help] How I can change the in_data 
in CustomOp
URL: 
https://github.com/apache/incubator-mxnet/issues/10294#issuecomment-386135871
 
 
   @wangshuailong please direct your questions to discuss.mxnet.io
   
   @nswamy @sandeep-krishnamurthy please label this - "Question"


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] anirudhacharya commented on issue #10290: c++ how to predict for a couple of images

[GitBox]

anirudhacharya commented on issue #10290: c++ how to predict for a couple of 
images
URL: 
https://github.com/apache/incubator-mxnet/issues/10290#issuecomment-386133435
 
 
   @nswamy @sandeep-krishnamurthy Please label this - "Question" and close this 
issue.


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] anirudhacharya commented on issue #10271: MxNet S3 driver sometimes fails to read even after 50 retries

[GitBox]

anirudhacharya commented on issue #10271: MxNet S3 driver sometimes fails to 
read even after 50 retries 
URL: 
https://github.com/apache/incubator-mxnet/issues/10271#issuecomment-386133118
 
 
   @ajayvohra2005 Please verify if the above commit fixed you problem.
   
   @nswamy @sandeep-krishnamurthy Please label this - "Pending Requester Info", 
"Bug".
   
   


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] anirudhacharya commented on issue #10264: Run rnn backward fail

[GitBox]

anirudhacharya commented on issue #10264: Run rnn backward fail 
URL: 
https://github.com/apache/incubator-mxnet/issues/10264#issuecomment-386132357
 
 
   @nswamy @sandeep-krishnamurthy Please label this - "Python", "RNN", "Bug".


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] anirudhacharya commented on issue #10357: Write Inplace isn't enabled in activation forward during inference.

[GitBox]

anirudhacharya commented on issue #10357: Write Inplace isn't enabled in 
activation forward during inference.
URL: 
https://github.com/apache/incubator-mxnet/issues/10357#issuecomment-386130562
 
 
   @zheng-da Could you please label this issue appropriately.


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] anirudhacharya commented on issue #10334: How to subtract mean during prediction with pretrained model (python)

[GitBox]

anirudhacharya commented on issue #10334: How to subtract mean during 
prediction with pretrained model (python)
URL: 
https://github.com/apache/incubator-mxnet/issues/10334#issuecomment-386129057
 
 
   @nswamy @sandeep-krishnamurthy Please label this - "Python", "Example"


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] anirudhacharya commented on issue #10329: Can't print params on gpu with collect_params

[GitBox]

anirudhacharya commented on issue #10329: Can't print params on gpu with 
collect_params
URL: 
https://github.com/apache/incubator-mxnet/issues/10329#issuecomment-386125785
 
 
   This issue is not reproducible with mxnet v1.1. Ran the following code -
   ```python
   import mxnet as mx
   from mxnet import gluon
   ctx = mx.gpu()
   net = gluon.nn.Sequential()
   with net.name_scope():
 net.add(gluon.rnn.LSTM(3, 1))
   net.collect_params().initialize(mx.init.Xavier(), ctx=ctx)
   params = net.collect_params()
   print(params)
   ```
   
   @nswamy @sandeep-krishnamurthy Please label this - "Gluon". And close the 
issue.


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] anirudh2290 commented on issue #8219: Broadcasting ops are slow

[GitBox]

anirudh2290 commented on issue #8219: Broadcasting ops are slow
URL: 
https://github.com/apache/incubator-mxnet/issues/8219#issuecomment-386123343
 
 
   Tracking here: https://issues.apache.org/jira/browse/MXNET-323


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] anirudhacharya commented on issue #10319: Hash functions in mxnet?

[GitBox]

anirudhacharya commented on issue #10319: Hash functions in mxnet?
URL: 
https://github.com/apache/incubator-mxnet/issues/10319#issuecomment-386122460
 
 
   @nswamy @sandeep-krishnamurthy Please label this - "Feature Request".


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] lanking520 commented on a change in pull request #10660: [MXNET-357] New Scala API Design (Symbol)

[GitBox]

lanking520 commented on a change in pull request #10660: [MXNET-357] New Scala 
API Design (Symbol)
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r185639573
 
 

 ##
 File path: 
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
 ##
 @@ -21,94 +21,101 @@ import scala.annotation.StaticAnnotation
 import scala.collection.mutable.ListBuffer
 import scala.language.experimental.macros
 import scala.reflect.macros.blackbox
-
 import org.apache.mxnet.init.Base._
 import org.apache.mxnet.utils.OperatorBuildUtils
 
 private[mxnet] class AddSymbolFunctions(isContrib: Boolean) extends 
StaticAnnotation {
   private[mxnet] def macroTransform(annottees: Any*) = macro 
SymbolImplMacros.addDefs
 }
 
+private[mxnet] class AddNewSymbolFunctions(isContrib: Boolean) extends 
StaticAnnotation {
+  private[mxnet] def macroTransform(annottees: Any*) = macro 
SymbolImplMacros.addNewDefs
+}
+
 private[mxnet] object SymbolImplMacros {
-  case class SymbolFunction(handle: SymbolHandle, keyVarNumArgs: String)
+  case class SymbolArg(argName: String, argType: String, isOptional : Boolean)
+  case class SymbolFunction(name: String, listOfArgs: List[SymbolArg])
 
   // scalastyle:off havetype
   def addDefs(c: blackbox.Context)(annottees: c.Expr[Any]*) = {
-impl(c)(false, annottees: _*)
+impl(c)(false, false, annottees: _*)
   }
-  // scalastyle:off havetype
+  def addNewDefs(c: blackbox.Context)(annottees: c.Expr[Any]*) = {
+impl(c)(false, true, annottees: _*)
 
 Review comment:
   I think we should keep using impl to avoid duplicated code, since the new 
API implementation is just a small component in


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] anirudhacharya commented on issue #10272: MxNet dmlc error that goes away on retry

[GitBox]

anirudhacharya commented on issue #10272: MxNet dmlc error that goes away on 
retry 
URL: 
https://github.com/apache/incubator-mxnet/issues/10272#issuecomment-386121552
 
 
   @ajayvohra2005 Please provide Environment Information.
   
   @nswamy @sandeep-krishnamurthy Please label this - "Build", "Flaky", 
"Pending Requester Info"
   
   


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] lanking520 commented on a change in pull request #10660: [MXNET-357] New Scala API Design (Symbol)

[GitBox]

lanking520 commented on a change in pull request #10660: [MXNET-357] New Scala 
API Design (Symbol)
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r185635887
 
 

 ##
 File path: 
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
 ##
 @@ -21,94 +21,101 @@ import scala.annotation.StaticAnnotation
 import scala.collection.mutable.ListBuffer
 import scala.language.experimental.macros
 import scala.reflect.macros.blackbox
-
 import org.apache.mxnet.init.Base._
 import org.apache.mxnet.utils.OperatorBuildUtils
 
 private[mxnet] class AddSymbolFunctions(isContrib: Boolean) extends 
StaticAnnotation {
   private[mxnet] def macroTransform(annottees: Any*) = macro 
SymbolImplMacros.addDefs
 }
 
+private[mxnet] class AddNewSymbolFunctions(isContrib: Boolean) extends 
StaticAnnotation {
+  private[mxnet] def macroTransform(annottees: Any*) = macro 
SymbolImplMacros.addNewDefs
+}
+
 private[mxnet] object SymbolImplMacros {
-  case class SymbolFunction(handle: SymbolHandle, keyVarNumArgs: String)
+  case class SymbolArg(argName: String, argType: String, isOptional : Boolean)
+  case class SymbolFunction(name: String, listOfArgs: List[SymbolArg])
 
   // scalastyle:off havetype
   def addDefs(c: blackbox.Context)(annottees: c.Expr[Any]*) = {
-impl(c)(false, annottees: _*)
+impl(c)(false, false, annottees: _*)
   }
-  // scalastyle:off havetype
+  def addNewDefs(c: blackbox.Context)(annottees: c.Expr[Any]*) = {
+impl(c)(false, true, annottees: _*)
+  }
+  // scalastyle:on havetype
 
-  private val symbolFunctions: Map[String, SymbolFunction] = initSymbolModule()
+  private val symbolFunctions: List[SymbolFunction] = initSymbolModule()
 
-  private def impl(c: blackbox.Context)(addSuper: Boolean, annottees: 
c.Expr[Any]*): c.Expr[Any] = {
+  private def impl(c: blackbox.Context)(addSuper: Boolean,
+newAPI: Boolean, annottees: 
c.Expr[Any]*): c.Expr[Any] = {
 import c.universe._
 
 val isContrib: Boolean = c.prefix.tree match {
   case q"new AddSymbolFunctions($b)" => c.eval[Boolean](c.Expr(b))
+  case q"new AddNewSymbolFunctions($b)" => c.eval[Boolean](c.Expr(b))
 }
 
 val newSymbolFunctions = {
-  if (isContrib) symbolFunctions.filter(_._1.startsWith("_contrib_"))
-  else symbolFunctions.filter(!_._1.startsWith("_contrib_"))
+  if (isContrib) symbolFunctions.filter(_.name.startsWith("_contrib_"))
+  else symbolFunctions.filter(!_.name.startsWith("_contrib_"))
 }
 
-val AST_TYPE_MAP_STRING_ANY = AppliedTypeTree(Ident(TypeName("Map")),
-  List(Ident(TypeName("String")), Ident(TypeName("Any"
-val AST_TYPE_MAP_STRING_STRING = AppliedTypeTree(Ident(TypeName("Map")),
-  List(Ident(TypeName("String")), Ident(TypeName("String"
-val AST_TYPE_SYMBOL_VARARG = AppliedTypeTree(
-  Select(
-Select(Ident(termNames.ROOTPKG), TermName("scala")),
-TypeName("")
-  ),
-  List(Select(Select(Select(
-Ident(TermName("org")), TermName("apache")), TermName("mxnet")), 
TypeName("Symbol")))
-)
-
-val functionDefs = newSymbolFunctions map { case (funcName, funcProp) =>
-  val functionScope = {
-if (isContrib) Modifiers()
-else {
-  if (funcName.startsWith("_")) Modifiers(Flag.PRIVATE) else 
Modifiers()
-}
+var functionDefs = List[DefDef]()
+
+if (!newAPI) {
+  functionDefs = newSymbolFunctions map { symbolfunction =>
+val funcName = symbolfunction.name
+val tName = TermName(funcName)
+q"""
+def $tName(name : String = null, attr : Map[String, String] = null)
+(args : org.apache.mxnet.Symbol*)(kwargs : Map[String, Any] = null)
+ : org.apache.mxnet.Symbol = {
+  createSymbolGeneral($funcName,name,attr,args,kwargs)
+  }
+ """.asInstanceOf[DefDef]
   }
-  val newName = {
-if (isContrib) funcName.substring(funcName.indexOf("_contrib_") + 
"_contrib_".length())
-else funcName
+} else {
+  functionDefs = newSymbolFunctions map { symbolfunction =>
+
+// Construct argument field
+var argDef = ListBuffer[String]()
+symbolfunction.listOfArgs.foreach(symbolarg => {
+  val currArgName = if (symbolarg.argName.equals("var")) "vari" else 
symbolarg.argName
+  if (symbolarg.isOptional) {
+argDef += s"${currArgName} : Option[${symbolarg.argType}] = None"
 
 Review comment:
   If we pass them as None, the arg will go with their default value in C


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] anirudhacharya commented on issue #10267: Continue train in R: what is begin_round?

[GitBox]

anirudhacharya commented on issue #10267: Continue train in R: what is 
begin_round?
URL: 
https://github.com/apache/incubator-mxnet/issues/10267#issuecomment-386120875
 
 
   @nswamy @sandeep-krishnamurthy Please label this - "R", "Bug".


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] anirudhacharya commented on issue #10295: Build failure with cython 0.28

[GitBox]

anirudhacharya commented on issue #10295: Build failure with cython 0.28
URL: 
https://github.com/apache/incubator-mxnet/issues/10295#issuecomment-386118541
 
 
   @nswamy @sandeep-krishnamurthy please label this - "Build"


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] roywei commented on issue #10775: RGB to gray ops

[GitBox]

roywei commented on issue #10775: RGB to gray ops
URL: 
https://github.com/apache/incubator-mxnet/issues/10775#issuecomment-386117526
 
 
   Hi @pribadihcr , please refer to the[ Image 
API](https://mxnet.incubator.apache.org/api/python/image/image.html), you can 
try `imdecode` with arg `to_rgb=0`. 
   could you also close this if it answers your question? 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] ashokei commented on issue #10591: [MXNET-365] handle inplace in mkldnn FallBackCompute

[GitBox]

ashokei commented on issue #10591: [MXNET-365] handle inplace in mkldnn 
FallBackCompute
URL: https://github.com/apache/incubator-mxnet/pull/10591#issuecomment-386117383
 
 
   @piiswrong can you please merge, 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] ashokei commented on issue #10591: [MXNET-365] handle inplace in mkldnn FallBackCompute

[GitBox]

ashokei commented on issue #10591: [MXNET-365] handle inplace in mkldnn 
FallBackCompute
URL: https://github.com/apache/incubator-mxnet/pull/10591#issuecomment-385096198
 
 
   @piiswrong can you please merge, 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] anirudhacharya commented on issue #10330: [Post 1.1][WIP] Couple of License Issues from 1.1 Release

[GitBox]

anirudhacharya commented on issue #10330: [Post 1.1][WIP] Couple of License 
Issues from 1.1 Release
URL: 
https://github.com/apache/incubator-mxnet/issues/10330#issuecomment-386114266
 
 
   @nswamy @sandeep-krishnamurthy Please label this - "Licenses"


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] anirudhacharya commented on issue #10330: [Post 1.1][WIP] Couple of License Issues from 1.1 Release

[GitBox]

anirudhacharya commented on issue #10330: [Post 1.1][WIP] Couple of License 
Issues from 1.1 Release
URL: 
https://github.com/apache/incubator-mxnet/issues/10330#issuecomment-386114266
 
 
   @apache/mxnet-committers Please label this - "Licenses"


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] lutzroeder closed issue #10359: Correct way to detect JSON file contains MXNet model

[GitBox]

lutzroeder closed issue #10359: Correct way to detect JSON file contains MXNet 
model
URL: https://github.com/apache/incubator-mxnet/issues/10359
 
 
   


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] anirudhacharya commented on issue #10359: Correct way to detect JSON file contains MXNet model

[GitBox]

anirudhacharya commented on issue #10359: Correct way to detect JSON file 
contains MXNet model
URL: 
https://github.com/apache/incubator-mxnet/issues/10359#issuecomment-386112131
 
 
   @lutzroeder if this is resolved, please close the issue.


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] zheng-da commented on issue #10731: Fix a bug in getting MKLDNN memory

[GitBox]

zheng-da commented on issue #10731: Fix a bug in getting MKLDNN memory
URL: https://github.com/apache/incubator-mxnet/pull/10731#issuecomment-386108437
 
 
   @marcoabreu do you have more comments?


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: [MXNET-359] fix checks on convolution parameters in MKLDNN. (#10666)

[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 1420697  [MXNET-359] fix checks on convolution parameters in MKLDNN. 
(#10666)
1420697 is described below

commit 14206978f461364c53aaf1c787e2f268e2a94b00
Author: Da Zheng 
AuthorDate: Wed May 2 13:20:13 2018 -0700

[MXNET-359] fix checks on convolution parameters in MKLDNN. (#10666)

* fix check on tuples of conv.

* check params in (de)conv.

* rename.

* add messages.
---
 src/operator/nn/convolution.cc | 18 -
 src/operator/nn/deconvolution.cc   | 22 +-
 src/operator/nn/mkldnn/mkldnn_base-inl.h   | 10 +--
 src/operator/nn/mkldnn/mkldnn_convolution.cc   | 69 +-
 src/operator/nn/mkldnn/mkldnn_deconvolution.cc | 99 +-
 5 files changed, 107 insertions(+), 111 deletions(-)

diff --git a/src/operator/nn/convolution.cc b/src/operator/nn/convolution.cc
index 7fd8bbb..0e8a929 100644
--- a/src/operator/nn/convolution.cc
+++ b/src/operator/nn/convolution.cc
@@ -54,7 +54,8 @@ static void ConvolutionComputeExCPU(const nnvm::NodeAttrs& 
attrs,
 const std::vector& inputs,
 const std::vector& req,
 const std::vector& outputs) {
-  if (SupportMKLDNNConv(inputs[0])) {
+  const ConvolutionParam& params = nnvm::get(attrs.parsed);
+  if (SupportMKLDNNConv(params, inputs[0])) {
 MKLDNN_OPCHECK_INIT(false, outputs.size(), inputs, outputs);
 MKLDNNConvolutionForward(attrs, ctx, inputs, req, outputs);
 MKLDNN_OPCHECK_RUN(ConvolutionCompute, attrs, ctx, inputs, req, 
outputs);
@@ -68,7 +69,8 @@ static void ConvolutionGradComputeExCPU(const 
nnvm::NodeAttrs& attrs,
 const std::vector& inputs,
 const std::vector& req,
 const std::vector& outputs) {
-  if (SupportMKLDNNConv(inputs[0])) {
+  const ConvolutionParam& params = nnvm::get(attrs.parsed);
+  if (SupportMKLDNNConv(params, inputs[0])) {
 MKLDNN_OPCHECK_INIT(true, outputs.size(), inputs, outputs);
 MKLDNNConvolutionBackward(attrs, ctx, inputs, req, outputs);
 MKLDNN_OPCHECK_RUN(ConvolutionGradCompute, attrs, ctx, inputs, req, 
outputs);
@@ -363,6 +365,18 @@ static void ConvolutionParamParser(nnvm::NodeAttrs* attrs) 
{
 if (param_.dilate.ndim() == 0) param_.dilate = Shape3(1, 1, 1);
 if (param_.pad.ndim() == 0) param_.pad = Shape3(0, 0, 0);
   }
+  CHECK_EQ(param_.kernel.ndim(), param_.stride.ndim())
+<< "Stride must have the same number of dimensions with kernel_size,"
+<< "but kernel_size is set to " << param_.kernel << " while stride is "
+<< param_.stride;
+  CHECK_EQ(param_.kernel.ndim(), param_.dilate.ndim())
+<< "Dilate must have the same number of dimensions with kernel_size,"
+<< "but kernel_size is set to " << param_.kernel << " while dilate is "
+<< param_.dilate;
+  CHECK_EQ(param_.kernel.ndim(), param_.pad.ndim())
+<< "Padding must have the same number of dimensions with kernel_size,"
+<< "but kernel_size is set to " << param_.kernel << " while padding is "
+<< param_.pad;
   attrs->parsed = std::move(param_);
 }
 
diff --git a/src/operator/nn/deconvolution.cc b/src/operator/nn/deconvolution.cc
index 0d1b391..13fc757 100644
--- a/src/operator/nn/deconvolution.cc
+++ b/src/operator/nn/deconvolution.cc
@@ -304,7 +304,8 @@ static void DeconvolutionComputeExCPU(const 
nnvm::NodeAttrs& attrs,
   const std::vector& inputs,
   const std::vector& req,
   const std::vector& outputs) {
-  if (SupportMKLDNNConv(inputs[0])) {
+  const DeconvolutionParam& param = 
nnvm::get(attrs.parsed);
+  if (SupportMKLDNNDeconv(param, inputs[0])) {
 MKLDNN_OPCHECK_INIT(false, outputs.size(), inputs, outputs);
 MKLDNNDeconvolutionForward(attrs, ctx, inputs, req, outputs);
 MKLDNN_OPCHECK_RUN(DeconvolutionCompute, attrs, ctx, inputs, req,
@@ -320,7 +321,8 @@ static void DeconvolutionGradComputeExCPU(const 
nnvm::NodeAttrs& attrs,
   const std::vector& inputs,
   const std::vector& req,
   const std::vector& outputs) 
{
-  if (SupportMKLDNNConv(inputs[0])) {
+  const DeconvolutionParam& param = 
nnvm::get(attrs.parsed);
+  if (SupportMKLDNNDeconv(param, inputs[0])) {
 MKLDNN_OPCHECK_INIT(true, outputs.size(), inputs, outputs);
 MKLDNNDeconvolutionBackward(attrs, ctx, inputs, req, outputs);
 MKLDNN_OPCHECK_RUN(DeconvolutionGradCompute, attrs, ctx, inputs, req,
@@ 

[GitHub] piiswrong closed pull request #10666: [MXNET-359] fix checks on convolution parameters in MKLDNN.

[GitBox]

piiswrong closed pull request #10666: [MXNET-359] fix checks on convolution 
parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666
 
 
   

This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:

As this is a foreign pull request (from a fork), the diff is supplied
below (as it won't show otherwise due to GitHub magic):

diff --git a/src/operator/nn/convolution.cc b/src/operator/nn/convolution.cc
index 7fd8bbb5599..0e8a929e1ba 100644
--- a/src/operator/nn/convolution.cc
+++ b/src/operator/nn/convolution.cc
@@ -54,7 +54,8 @@ static void ConvolutionComputeExCPU(const nnvm::NodeAttrs& 
attrs,
 const std::vector& inputs,
 const std::vector& req,
 const std::vector& outputs) {
-  if (SupportMKLDNNConv(inputs[0])) {
+  const ConvolutionParam& params = nnvm::get(attrs.parsed);
+  if (SupportMKLDNNConv(params, inputs[0])) {
 MKLDNN_OPCHECK_INIT(false, outputs.size(), inputs, outputs);
 MKLDNNConvolutionForward(attrs, ctx, inputs, req, outputs);
 MKLDNN_OPCHECK_RUN(ConvolutionCompute, attrs, ctx, inputs, req, 
outputs);
@@ -68,7 +69,8 @@ static void ConvolutionGradComputeExCPU(const 
nnvm::NodeAttrs& attrs,
 const std::vector& inputs,
 const std::vector& req,
 const std::vector& outputs) {
-  if (SupportMKLDNNConv(inputs[0])) {
+  const ConvolutionParam& params = nnvm::get(attrs.parsed);
+  if (SupportMKLDNNConv(params, inputs[0])) {
 MKLDNN_OPCHECK_INIT(true, outputs.size(), inputs, outputs);
 MKLDNNConvolutionBackward(attrs, ctx, inputs, req, outputs);
 MKLDNN_OPCHECK_RUN(ConvolutionGradCompute, attrs, ctx, inputs, req, 
outputs);
@@ -363,6 +365,18 @@ static void ConvolutionParamParser(nnvm::NodeAttrs* attrs) 
{
 if (param_.dilate.ndim() == 0) param_.dilate = Shape3(1, 1, 1);
 if (param_.pad.ndim() == 0) param_.pad = Shape3(0, 0, 0);
   }
+  CHECK_EQ(param_.kernel.ndim(), param_.stride.ndim())
+<< "Stride must have the same number of dimensions with kernel_size,"
+<< "but kernel_size is set to " << param_.kernel << " while stride is "
+<< param_.stride;
+  CHECK_EQ(param_.kernel.ndim(), param_.dilate.ndim())
+<< "Dilate must have the same number of dimensions with kernel_size,"
+<< "but kernel_size is set to " << param_.kernel << " while dilate is "
+<< param_.dilate;
+  CHECK_EQ(param_.kernel.ndim(), param_.pad.ndim())
+<< "Padding must have the same number of dimensions with kernel_size,"
+<< "but kernel_size is set to " << param_.kernel << " while padding is "
+<< param_.pad;
   attrs->parsed = std::move(param_);
 }
 
diff --git a/src/operator/nn/deconvolution.cc b/src/operator/nn/deconvolution.cc
index 0d1b391104a..13fc757fb2d 100644
--- a/src/operator/nn/deconvolution.cc
+++ b/src/operator/nn/deconvolution.cc
@@ -304,7 +304,8 @@ static void DeconvolutionComputeExCPU(const 
nnvm::NodeAttrs& attrs,
   const std::vector& inputs,
   const std::vector& req,
   const std::vector& outputs) {
-  if (SupportMKLDNNConv(inputs[0])) {
+  const DeconvolutionParam& param = 
nnvm::get(attrs.parsed);
+  if (SupportMKLDNNDeconv(param, inputs[0])) {
 MKLDNN_OPCHECK_INIT(false, outputs.size(), inputs, outputs);
 MKLDNNDeconvolutionForward(attrs, ctx, inputs, req, outputs);
 MKLDNN_OPCHECK_RUN(DeconvolutionCompute, attrs, ctx, inputs, req,
@@ -320,7 +321,8 @@ static void DeconvolutionGradComputeExCPU(const 
nnvm::NodeAttrs& attrs,
   const std::vector& inputs,
   const std::vector& req,
   const std::vector& outputs) 
{
-  if (SupportMKLDNNConv(inputs[0])) {
+  const DeconvolutionParam& param = 
nnvm::get(attrs.parsed);
+  if (SupportMKLDNNDeconv(param, inputs[0])) {
 MKLDNN_OPCHECK_INIT(true, outputs.size(), inputs, outputs);
 MKLDNNDeconvolutionBackward(attrs, ctx, inputs, req, outputs);
 MKLDNN_OPCHECK_RUN(DeconvolutionGradCompute, attrs, ctx, inputs, req,
@@ -356,6 +358,22 @@ static void DeconvolutionParamParser(nnvm::NodeAttrs* 
attrs) {
 if (param_.pad.ndim() == 0) param_.pad = Shape3(0, 0, 0);
 if (param_.adj.ndim() == 0) param_.adj = Shape3(0, 0, 0);
   }
+  CHECK_EQ(param_.kernel.ndim(), param_.stride.ndim())
+<< "Stride must have the same number of dimensions with kernel_size,"
+<< "but kernel_size is set to " << param_.kernel << " while stride is "
+<< param_.stride;
+  CHECK_EQ(param_.kernel.ndim(), param_.dilate.ndim())
+<< "Dilate must have the same number of dimensions with kernel_size,"
+<< "but kernel_size is set to " << param_.kernel 

[GitHub] aaronmarkham commented on issue #10728: fixed broken links and added jenkins info

[GitBox]

aaronmarkham commented on issue #10728: fixed broken links and added jenkins 
info
URL: https://github.com/apache/incubator-mxnet/pull/10728#issuecomment-386094012
 
 
   @ThomasDelteil you're right. I made one in build_version_doc that is up to 
date. 
   The one in docs was for folks that wanted to build locally and for just 
their current fork/branch, but is was/is broken.


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] lanking520 commented on a change in pull request #10660: [MXNET-357] New Scala API Design (Symbol)

[GitBox]

lanking520 commented on a change in pull request #10660: [MXNET-357] New Scala 
API Design (Symbol)
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r185611194
 
 

 ##
 File path: 
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
 ##
 @@ -136,20 +143,70 @@ private[mxnet] object SymbolImplMacros {
 result
   }
 
+  // Convert C++ Types to Scala Types
+  private def typeConversion(in : String, argType : String = "") : String = {
+in match {
+  case "Shape(tuple)" | "ShapeorNone" => "org.apache.mxnet.Shape"
+  case "Symbol" | "NDArray" | "NDArray-or-Symbol" => 
"org.apache.mxnet.Symbol"
+  case "Symbol[]" | "NDArray[]" | "NDArray-or-Symbol[]" | 
"SymbolorSymbol[]"
+  => "Array[org.apache.mxnet.Symbol]"
+  case "float" | "real_t" | "floatorNone" => 
"org.apache.mxnet.Base.MXFloat"
+  case "int" | "intorNone" | "int(non-negative)" => "Int"
+  case "long" | "long(non-negative)" => "Long"
+  case "double" => "Double"
+  case "string" => "String"
+  case "boolean" => "Boolean"
+  case "tupleof" => "Any"
+  case default => throw new IllegalArgumentException(
+s"Invalid type for args: $default, $argType")
+}
+  }
+
+
+
+  private def argumentCleaner(argType : String) : (String, Boolean) = {
+val spaceRemoved = argType.replaceAll("\\s+", "")
+var commaRemoved : Array[String] = new Array[String](0)
+// Deal with the case e.g: stype : {'csr', 'default', 'row_sparse'}
+if (spaceRemoved.charAt(0)== '{') {
+  val endIdx = spaceRemoved.indexOf('}')
+  commaRemoved = spaceRemoved.substring(endIdx + 1).split(",")
+  commaRemoved(0) = "string"
+} else {
+  commaRemoved = spaceRemoved.split(",")
+}
+// Optional Field
+if (commaRemoved.length >= 3) {
+  // arg: Type, optional, default = Null
+  require(commaRemoved(1).equals("optional"))
+  require(commaRemoved(2).startsWith("default="))
+  (typeConversion(commaRemoved(0), argType), true)
+} else if (commaRemoved.length == 2 || commaRemoved.length == 1) {
+  val tempType = typeConversion(commaRemoved(0), argType)
+  val tempOptional = tempType.equals("org.apache.mxnet.Symbol")
+  (tempType, tempOptional)
+} else {
+  throw new IllegalArgumentException(
+s"Unrecognized arg field: $argType, ${commaRemoved.length}")
+}
+
+  }
+
+
   // List and add all the atomic symbol functions to current module.
-  private def initSymbolModule(): Map[String, SymbolFunction] = {
+  private def initSymbolModule(): List[SymbolFunction] = {
 val opNames = ListBuffer.empty[String]
 _LIB.mxListAllOpNames(opNames)
-opNames.map(opName => {
+opNames.filter(op => !op.startsWith("_") || 
op.startsWith("_contrib_")).map(opName => {
 
 Review comment:
   Sure, let's add a TODO in here to make sure adding more functions.


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] lanking520 commented on a change in pull request #10660: [MXNET-357] New Scala API Design (Symbol)

[GitBox]

lanking520 commented on a change in pull request #10660: [MXNET-357] New Scala 
API Design (Symbol)
URL: https://github.com/apache/incubator-mxnet/pull/10660#discussion_r185610879
 
 

 ##
 File path: 
scala-package/macros/src/main/scala/org/apache/mxnet/SymbolMacro.scala
 ##
 @@ -136,20 +143,70 @@ private[mxnet] object SymbolImplMacros {
 result
   }
 
+  // Convert C++ Types to Scala Types
+  private def typeConversion(in : String, argType : String = "") : String = {
+in match {
+  case "Shape(tuple)" | "ShapeorNone" => "org.apache.mxnet.Shape"
+  case "Symbol" | "NDArray" | "NDArray-or-Symbol" => 
"org.apache.mxnet.Symbol"
 
 Review comment:
   Yes, they share the same Function name


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

svn commit: r26635 - in /dev/incubator/mxnet/1.2.0.rc2: ./ apache-mxnet-src-1.2.0.rc2-incubating.tar.gz apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.asc apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.md

[anirudh2290]

Author: anirudh2290
Date: Wed May  2 19:13:13 2018
New Revision: 26635

Log:
Add mxnet-1.2.0.rc2

Added:
dev/incubator/mxnet/1.2.0.rc2/
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz  
 (with props)

dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.asc

dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.md5

dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.sha512

Added: 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz
==
Binary file - no diff available.

Propchange: 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz
--
svn:mime-type = application/octet-stream

Added: 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.asc
==
--- 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.asc 
(added)
+++ 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.asc 
Wed May  2 19:13:13 2018
@@ -0,0 +1,11 @@
+-BEGIN PGP SIGNATURE-
+Version: GnuPG v1
+
+iQEcBAABAgAGBQJa6gzkAAoJEDjVoQyIJMHAujwH/i5UdTik6apawrSBTrWqjox+
+3mztGt/0UacX20bz57gWdlIAAeKSLYUic6jL7noKALYInmR/rC64HAf7en86R8JJ
+VwZNtNG72O99hGd7M2HkGomAE8ZZ10hG6vvINiOZeyK1VMdvraCOBxS8/xevQ8c1
+BfYTsRcKAPDjXn1hQ5sHBCZTRSAZju+G+JN6HpSqhYsqojHQYGr4vnzlSIoCuzPB
+WlXeCNloVrF8lRVLnlwl7+yhM97aMzjHJ72vZjg4aA0ahNx6sxEgYi8PFJ0aE04V
+/SO8sZSDj6X/U1wznL6rERN+uVEF66pytvnSuV2rAq37eUbQm5n173ubgpiVOog=
+=hAln
+-END PGP SIGNATURE-

Added: 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.md5
==
--- 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.md5 
(added)
+++ 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.md5 
Wed May  2 19:13:13 2018
@@ -0,0 +1 @@
+90843c449ea8cc27ff4554bd9b388d6f  apache-mxnet-src-1.2.0.rc2-incubating.tar.gz

Added: 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.sha512
==
--- 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.sha512
 (added)
+++ 
dev/incubator/mxnet/1.2.0.rc2/apache-mxnet-src-1.2.0.rc2-incubating.tar.gz.sha512
 Wed May  2 19:13:13 2018
@@ -0,0 +1 @@
+8e26515a77f6b7bdd8448d89af484ae6b441500e900c6c08e68f589669315e22a59d3785ecd73845e577f438be4004a2a2cae5d391917c76934136e38134b208
  apache-mxnet-src-1.2.0.rc2-incubating.tar.gz

[GitHub] aaronmarkham commented on issue #10785: clarify docs build

[GitBox]

aaronmarkham commented on issue #10785: clarify docs build
URL: https://github.com/apache/incubator-mxnet/pull/10785#issuecomment-386079237
 
 
   @kpmurali 


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] ashokei commented on issue #10736: [MXNET-367] update mkldnn to v0.14 and disable building test examples

[GitBox]

ashokei commented on issue #10736: [MXNET-367] update mkldnn to v0.14 and 
disable building test examples
URL: https://github.com/apache/incubator-mxnet/pull/10736#issuecomment-386072794
 
 
   @marcoabreu updated all locations, can you please merge if ok. 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 #10666: [MXNET-359] fix checks on convolution parameters in MKLDNN.

[GitBox]

piiswrong commented on a change in pull request #10666: [MXNET-359] fix checks 
on convolution parameters in MKLDNN.
URL: https://github.com/apache/incubator-mxnet/pull/10666#discussion_r185582075
 
 

 ##
 File path: src/operator/nn/convolution.cc
 ##
 @@ -363,6 +365,9 @@ static void ConvolutionParamParser(nnvm::NodeAttrs* attrs) 
{
 if (param_.dilate.ndim() == 0) param_.dilate = Shape3(1, 1, 1);
 if (param_.pad.ndim() == 0) param_.pad = Shape3(0, 0, 0);
   }
+  CHECK_EQ(param_.kernel.ndim(), param_.stride.ndim());
 
 Review comment:
   These checks needs to have error messages.
   "stride must have the same number of dimensions with kernel_size, but 
kernel_size is set to (x,x,x) while stride is (x,x)"


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 #10761: update the performance page of MXNet.

[GitBox]

piiswrong closed pull request #10761: update the performance page of MXNet.
URL: https://github.com/apache/incubator-mxnet/pull/10761
 
 
   

This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:

As this is a foreign pull request (from a fork), the diff is supplied
below (as it won't show otherwise due to GitHub magic):

diff --git a/docs/faq/perf.md b/docs/faq/perf.md
index b5d73f69a03..ce74391122d 100644
--- a/docs/faq/perf.md
+++ b/docs/faq/perf.md
@@ -29,65 +29,70 @@ Note that _MXNet_ treats all CPUs on a single machine as a 
single device.
 So whether you specify `cpu(0)` or `cpu()`, _MXNet_ will use all CPU cores on 
the machine.
 
 ### Scoring results
-The following table shows performance,
+The following table shows performance of 
[MXNet-1.2.0.rc1](https://github.com/apache/incubator-mxnet/releases/download/1.2.0.rc1/apache-mxnet-src-1.2.0.rc1-incubating.tar.gz),
 namely number of images that can be predicted per second.
 We used 
[example/image-classification/benchmark_score.py](https://github.com/dmlc/mxnet/blob/master/example/image-classification/benchmark_score.py)
 to measure the performance on different AWS EC2 machines.
 
-AWS EC2 C4.8xlarge:
-
-| Batch | Alexnet | VGG | Inception-BN | Inception-v3 | Resnet 50 | Resnet 152 
|
-| --- | --- | --- | --- | --- | --- | --- |
-|   1 |  119.57 | 34.23 |  111.36 |  54.42 |  42.83 | 19.51 |
-|   2 | 210.58 | 51.63 |  137.10 |  67.30 |  57.54 | 23.56 |
-|   4 | 318.54 | 70.00 |  187.21 |  76.53 |  63.64 | 25.80 |
-|   8 | 389.34 | 77.39 |  211.90 |  84.26 |  63.89 | 28.11 |
-|  16 | 489.12 | 85.26 |  220.52 |  82.00 |  63.93 | 27.08 |
-|  32 | 564.04 | 87.15 |  208.21 |  83.05 |  62.19 | 25.76 |
-
-AWS EC2 C4.4xlarge:
-
-| Batch | Alexnet | VGG | Inception-BN | Inception-v3 | Resnet 50 | Resnet 152 
|
-| --- | --- | --- | --- | --- | --- | --- |
-|   1 |  109.96 | 23.00 |  71.82 |  28.10 |  30.66 | 11.81 |
-|   2 | 124.56 | 24.86 |  81.61 |  31.32 |  32.73 | 12.82 |
-|   4 | 157.01 | 26.60 |  86.77 |  32.94 |  33.32 | 13.16 |
-|   8 | 178.40 | 30.67 |  88.58 |  33.52 |  33.32 | 13.32 |
-|  16 | 189.52 | 35.61 |  90.36 |  33.63 |  32.94 | 13.18 |
-|  32 | 196.61 | 38.98 |  105.27 |  33.77 |  32.65 | 13.00 |
-
-AWS EC2 C4.2xlarge:
-
-| Batch | Alexnet | VGG | Inception-BN | Inception-v3 | Resnet 50 | Resnet 152 
|
-| --- | --- | --- | --- | --- | --- | --- |
-|   1 |  70.75 | 12.87 |  42.86 |  16.53 |  18.14 | 7.01 |
-|   2 | 71.53 | 13.08 |  45.66 |  17.38 |  18.53 | 7.18 |
-|   4 | 84.72 | 15.38 |  47.50 |  17.80 |  18.96 | 7.35 |
-|   8 | 93.44 | 18.33 |  48.08 |  17.93 |  18.99 | 7.40 |
-|  16 | 97.03 | 20.12 |  55.73 |  18.00 |  18.91 | 7.36 |
-|  32 | 113.90 | 21.10 |  62.54 |  17.98 |  18.80 | 7.33 |
-
-AWS EC2 C4.xlarge:
-
-| Batch | Alexnet | VGG | Inception-BN | Inception-v3 | Resnet 50 | Resnet 152 
|
-| --- | --- | --- | --- | --- | --- | --- |
-|   1 |  37.92 | 6.57 |  23.09 |  8.79 |  9.65 | 3.73 |
-|   2 | 36.77 | 7.31 |  24.00 |  9.00 |  9.84 | 3.78 |
-|   4 | 43.18 | 8.94 |  24.42 |  9.12 |  9.91 | 3.83 |
-|   8 | 47.05 | 10.01 |  28.32 |  9.13 |  9.88 | 3.83 |
-|  16 | 55.74 | 10.61 |  31.96 |  9.14 |  9.86 | 3.80 |
-|  32 | 65.05 | 10.91 |  33.86 |  9.34 |  10.31 | 3.86 |
-
-AWS EC2 C4.large:
-
-| Batch | Alexnet | VGG | Inception-BN | Inception-v3 | Resnet 50 | Resnet 152 
|
-| --- | --- | --- | --- | --- | --- | --- |
-|   1 |  19.86 | 3.67 |  12.20 |  4.59 |  5.11 | 1.97 |
-|   2 | 19.37 | 4.24 |  12.41 |  4.64 |  5.15 | 1.98 |
-|   4 | 22.64 | 4.89 |  14.34 |  4.66 |  5.16 | 2.00 |
-|   8 | 27.19 | 5.25 |  16.17 |  4.66 |  5.16 | 1.99 |
-|  16 | 31.82 | 5.46 |  17.24 |  4.76 |  5.35 | OOM |
-|  32 | 34.67 | 5.55 |  17.64 |  4.88 |  OOM | OOM |
+AWS EC2 C5.18xlarge:
+
+| Batch | Alexnet | VGG| Inception-BN | Inception-v3 | Resnet 50 | Resnet 
152 |
+|---|-||--|--|---||
+| 1 | 390.53  | 81.57  | 124.13   | 62.26| 76.22 | 32.92   
   |
+| 2 | 596.45  | 100.84 | 206.58   | 93.36| 119.55| 46.80   
   |
+| 4 | 710.77  | 119.04 | 275.55   | 127.86   | 148.62| 59.36   
   |
+| 8 | 921.40  | 120.38 | 380.82   | 157.11   | 167.95| 70.78   
   |
+| 16| 1018.43 | 115.30 | 411.67   | 168.71   | 178.54| 75.13   
   |
+| 32| 1290.31 | 107.19 | 483.34   | 179.38   | 193.47| 85.86   
   |
+
+
+AWS EC2 C5.9xlarge:
+
+| Batch | Alexnet | VGG   | Inception-BN | Inception-v3 | Resnet 50 | Resnet 
152 |
+|---|-|---|--|--|---||
+| 1 | 257.77  | 50.61 | 130.99   | 66.95| 75.38 | 32.33
  |
+| 2 | 410.60  | 63.02 | 195.14   | 87.84| 102.67| 41.57
  |
+| 4 | 462.59  | 62.64 | 263.15   | 109.87   | 127.15| 50.69
  |
+| 8 | 573.79  | 63.95 | 309.99   | 121.36   | 

[GitHub] zheng-da commented on a change in pull request #10783: Fix 'make clean USE_MKLDNN=1' will build mkldnn issue

[GitBox]

zheng-da commented on a change in pull request #10783: Fix 'make clean 
USE_MKLDNN=1' will build mkldnn issue
URL: https://github.com/apache/incubator-mxnet/pull/10783#discussion_r185574502
 
 

 ##
 File path: Makefile
 ##
 @@ -66,13 +66,22 @@ $(warning "USE_MKL2017 is deprecated. We will switch to 
USE_MKLDNN.")
 endif
 
 ifeq ($(USE_MKLDNN), 1)
+ifeq ($(MAKECMDGOALS),)
 
 Review comment:
   so now people need to define MAKECMDGOALS manually to compile mxnet with 
mkldnn?


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