[GitHub] chinakook closed issue #7517: Add Depthwise Deconvolution support?

2017-08-27 Thread git 
chinakook closed issue #7517: Add Depthwise Deconvolution support?
URL: https://github.com/apache/incubator-mxnet/issues/7517
 
 
   
 

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


With regards,
Apache Git Services

[GitHub] eric-haibin-lin opened a new pull request #7638: CSRNDArray from/to scipy csr_matrix; fix rand_shape_nd

2017-08-27 Thread git 
eric-haibin-lin opened a new pull request #7638: CSRNDArray from/to scipy 
csr_matrix; fix rand_shape_nd
URL: https://github.com/apache/incubator-mxnet/pull/7638
 
 
   Added preliminary/inefficient support so that user can 
   - construct CSRNDArray from scipy csr_matrix. 
   - convert CSRNDArray to scipy csr_matrix 
   
   Also fixed `rand_shape_nd` so that it returns tuple (consistent w/ 
rand_shape_2d)
   
   TODO:
   - move copy logic to cpp backend to reduce copy/blocking. 
   
   @reminisce @anirudh2290 
 

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] im9uri commented on issue #6245: Compile lastest MXNet failed with cude 8.0

2017-08-27 Thread git 
im9uri commented on issue #6245: Compile lastest MXNet failed with cude 8.0
URL: 
https://github.com/apache/incubator-mxnet/issues/6245#issuecomment-325259554
 
 
   @xinario I replaced every instance of CUDA_R_32I in MXNet with a type that 
was defined in the library_type.h header file. It was probably something along 
the line of CUDA_R_32F. Obviously not the best workaround... but the build 
passed.
 

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


With regards,
Apache Git Services

[GitHub] qingzhouzhen closed issue #7594: pvanet Incompatible input shape, why?

2017-08-27 Thread git 
qingzhouzhen closed issue #7594: pvanet Incompatible input shape, why?
URL: https://github.com/apache/incubator-mxnet/issues/7594
 
 
   
 

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 #7476: R-package RNN refactor

2017-08-27 Thread git 
jeremiedb commented on issue #7476: R-package RNN refactor
URL: https://github.com/apache/incubator-mxnet/pull/7476#issuecomment-325228128
 
 
   Improve the harmonization with model.FeedForward: same fixed.params, 
arg.params and aux.params input arguments. 
   
   Remove redundancies between model.buckets and model.train.buckets. 
   
   model.buckets is as performant on single symbol model as model.Feedforward. 
If using BatchNorm, single symbol works fine but the still is issue if training 
on list of symbols. 
   
   Comment out the lstm in testthat to pass CI.
   
   Integrate CPU compatible RNN construction with raw lstm and gru cells in 
rnn.graph. No support for masking. Still need to test efficiency of inference 
and potentially adapt the inference functions. 
   
   Unsure if wouldn't be preferable to assume a batch.size X seq.length input 
dimension to interator as it's the format expectged by symnol.RNN cell. Or add 
a shape detector to handle it automatically as in the python 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] atiyo opened a new issue #7637: Strange Validation and Training Losses at epoch change

2017-08-27 Thread git 
atiyo opened a new issue #7637: Strange Validation and Training Losses at epoch 
change
URL: https://github.com/apache/incubator-mxnet/issues/7637
 
 
   I struggled to get some mxnet models training to a good accuracy, so I took 
a closer look at training and validation losses of a toy model. I noticed some 
strange spikes between epochs, which surprised me.
   
   I expect I'm likely doing something wrong, but I can't see what: I have 
tried several optimisers, with learning rates spanning several orders of 
magnitude. It's most plausible that I'm doing something drastically wrong, 
being new to mxnet.
   
   Graphic below illustrating the phenomenon, and also code to reproduce figure:
   
   
![adam_loss](https://user-images.githubusercontent.com/12828061/29753519-35b2e6e2-8b6b-11e7-8c08-14b8730efceb.png)
   
   
   ```
   import mxnet as mx
   import numpy as np
   
   optimizer_choice = 'adam'
   learning_rate = 0.01
   batch_size = 500
   
   inputs = np.expand_dims(np.random.uniform(low=0., high=2*np.pi, size=1), 
axis=1)
   labels = np.sin(inputs)
   
   eval_inputs = np.expand_dims(np.random.uniform(low=0., high=2*np.pi, 
size=1), axis=1)
   eval_labels = np.sin(eval_inputs)
   
   data_iter = mx.io.NDArrayIter(data={'data':inputs}, label={'label':labels}, 
batch_size=batch_size, shuffle=True)
   eval_data_iter = mx.io.NDArrayIter(data={'data':eval_inputs}, 
label={'label':eval_labels}, batch_size=batch_size, shuffle=True)
   
   data = mx.sym.Variable('data')
   label = mx.sym.Variable('label')
   fc1 = mx.sym.FullyConnected(data=data, num_hidden=128)
   ac1 = mx.sym.Activation(data=fc1, act_type='relu')
   fc2 = mx.sym.FullyConnected(data=ac1, num_hidden=64)
   ac2 = mx.sym.Activation(data=fc2, act_type='relu')
   fc3 = mx.sym.FullyConnected(data=ac2, num_hidden=16)
   ac3 = mx.sym.Activation(data=fc3, act_type='relu')
   fc4 = mx.sym.FullyConnected(data=ac3, num_hidden=1)
   ac4 = mx.sym.Activation(data=fc4, act_type='tanh')
   loss = mx.symbol.LinearRegressionOutput(data=ac4, label=label)
   net = mx.module.Module(symbol=loss, data_names=['data'], 
label_names=['label'])
   
   train_error = [] 
   eval_error = []
   def log_error(period, log):
   def _callback(param):
   if param.nbatch % period == 0:
   name, value = param.eval_metric.get()
   log.append(value)
   return _callback
   
   optimizer_params={'learning_rate':learning_rate}
   net.fit(data_iter,
 optimizer=optimizer_choice,
 optimizer_params=optimizer_params,
 eval_data=eval_data_iter,
 eval_metric='mse',
 num_epoch=5,
 epoch_end_callback = mx.callback.do_checkpoint('test_net'),
 eval_batch_end_callback = log_error(1,eval_error),
 batch_end_callback = log_error(1,train_error)
 )
   
   train_error = np.array(train_error)
   eval_error = np.array(eval_error)
   import matplotlib.pyplot as plt
   plt.plot(np.arange(train_error.size),train_error, label = 'Training Error')
   plt.plot(np.arange(eval_error.size), eval_error, label = 'Validation Error')
   plt.legend(loc='upper right')
   plt.xlabel('Batch Number')
   plt.ylabel('Error')
   plt.title('Optimizer: {}. Learning Rate: 
{}'.format(optimizer_choice,learning_rate))
   plt.gca().set_ylim(bottom=0)
   plt.show()
   
   ```
   ## Environment info
   Operating System: macOS
   
   MXNet version: 0.11.0
   
   Python version and distribution: Python 2.7.13
   
   
   
   
 

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


With regards,
Apache Git Services

[GitHub] zhanghang1989 commented on issue #7579: style transfer example

2017-08-27 Thread git 
zhanghang1989 commented on issue #7579: style transfer example
URL: https://github.com/apache/incubator-mxnet/pull/7579#issuecomment-325035839
 
 
   I have sent a PR to web-data with the images 
https://github.com/dmlc/web-data/pull/26
   I will update the image paths after it is merged. @szha @piiswrong 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] chinakook commented on issue #7613: 1x1 convolution acceleration

2017-08-27 Thread git 
chinakook commented on issue #7613: 1x1 convolution acceleration
URL: https://github.com/apache/incubator-mxnet/pull/7613#issuecomment-325193542
 
 
   @reminisce I will complete the testings. I think performance improvement and 
memory reduction will show up in the image detection and segmentation cases 
where feature maps are very big(without cudnn). You can refer to the 
tensorflow's optimization for the 1x1 conv, it's more simple and more clear 
than Caffe:
   [tensorflow conv 
forward](https://github.com/tensorflow/tensorflow/blob/a0d784bdd31b27e013a7eac58a86ba62e86db299/tensorflow/core/kernels/conv_ops_using_gemm.cc#L238)
   [tensorflow conv 
backward](https://github.com/tensorflow/tensorflow/blob/42ca99b5aae03a8122ba0db94abfe1f3f5c257dc/tensorflow/core/kernels/conv_grad_input_ops.cc#L666)
 

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 commented on issue #7610: error C2768 "linalg_gemm":Illegal use of explicit template arguments

2017-08-27 Thread git 
chinakook commented on issue #7610: error C2768 "linalg_gemm":Illegal use of 
explicit template arguments
URL: 
https://github.com/apache/incubator-mxnet/issues/7610#issuecomment-325193766
 
 
   @piiswrong 
   Thanks, It's partly solved. New issues:
   d:\proj\dev\mx\src\operator\linalg_impl.h(442): error C2872: 'cpu': 
ambiguous symbol (compiling source file 
D:\proj\dev\mx\src\operator\upsampling.cc)
   d:\proj\dev\mx\src\operator\linalg_impl.h(443): error C2872: 'cpu': 
ambiguous symbol (compiling source file 
D:\proj\dev\mx\src\operator\upsampling.cc)
 

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 commented on issue #7610: error C2768 "linalg_gemm":Illegal use of explicit template arguments

2017-08-27 Thread git 
chinakook commented on issue #7610: error C2768 "linalg_gemm":Illegal use of 
explicit template arguments
URL: 
https://github.com/apache/incubator-mxnet/issues/7610#issuecomment-325193766
 
 
   It's partly solved. New issues:
   d:\proj\dev\mx\src\operator\linalg_impl.h(442): error C2872: 'cpu': 
ambiguous symbol (compiling source file 
D:\proj\dev\mx\src\operator\upsampling.cc)
   d:\proj\dev\mx\src\operator\linalg_impl.h(443): error C2872: 'cpu': 
ambiguous symbol (compiling source file 
D:\proj\dev\mx\src\operator\upsampling.cc)
 

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 commented on issue #7613: 1x1 convolution acceleration

2017-08-27 Thread git 
chinakook commented on issue #7613: 1x1 convolution acceleration
URL: https://github.com/apache/incubator-mxnet/pull/7613#issuecomment-325193542
 
 
   @reminisce I will complete the testings. I think performance improvement and 
memory reduction will show up in the image detection and segmentation cases 
where feature maps are very big. You can refer to the tensorflow's optimization 
for the 1x1 conv, it's more simple and more clear than Caffe:
   [tensorflow conv 
forward](https://github.com/tensorflow/tensorflow/blob/a0d784bdd31b27e013a7eac58a86ba62e86db299/tensorflow/core/kernels/conv_ops_using_gemm.cc#L238)
   [tensorflow conv 
backward](https://github.com/tensorflow/tensorflow/blob/42ca99b5aae03a8122ba0db94abfe1f3f5c257dc/tensorflow/core/kernels/conv_grad_input_ops.cc#L666)
 

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: [build] explicitly install JDK8 (#7574)

2017-08-27 Thread liuyizhi 
This is an automated email from the ASF dual-hosted git repository.

liuyizhi 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 e051297  [build] explicitly install JDK8 (#7574)
e051297 is described below

commit e05129774e76206fe890b511c346953107b05fce
Author: Nan Zhu 
AuthorDate: Sun Aug 27 01:12:23 2017 -0700

[build] explicitly install JDK8 (#7574)

* explicitly install openjdk8

* handle earlier version of ubuntu

* install software-properties-common

* update -y

* update commands
---
 docker/install/scala.sh| 10 +-
 docs/get_started/build_from_source.md  |  8 +++-
 tests/ci_build/Dockerfile.ubuntu1404_cuda75_cudnn5 |  8 +++-
 tests/ci_build/install/ubuntu_install_scala.sh |  9 +++--
 4 files changed, 30 insertions(+), 5 deletions(-)

diff --git a/docker/install/scala.sh b/docker/install/scala.sh
index bb0bb9c..c1d2de6 100755
--- a/docker/install/scala.sh
+++ b/docker/install/scala.sh
@@ -19,7 +19,15 @@
 
 # install libraries for mxnet's scala package on ubuntu
 
-apt-get install -y maven default-jdk
+
+apt-get install -y software-properties-common
+add-apt-repository -y ppa:webupd8team/java
+apt-get update
+echo "oracle-java8-installer shared/accepted-oracle-license-v1-1 select true" 
| debconf-set-selections
+apt-get install -y oracle-java8-installer
+apt-get install -y oracle-java8-set-default
+
+apt-get install -y maven 
 
 wget http://downloads.lightbend.com/scala/2.11.8/scala-2.11.8.deb
 dpkg -i scala-2.11.8.deb
diff --git a/docs/get_started/build_from_source.md 
b/docs/get_started/build_from_source.md
index 4ff2cc0..9bf397b 100644
--- a/docs/get_started/build_from_source.md
+++ b/docs/get_started/build_from_source.md
@@ -367,7 +367,13 @@ Both JDK and Maven are required to build the Scala package.
 
 
 ```bash
-sudo apt-get install -y maven default-jdk
+apt-get install -y software-properties-common
+add-apt-repository -y ppa:webupd8team/java
+apt-get update
+echo "oracle-java8-installer shared/accepted-oracle-license-v1-1 select true" 
| debconf-set-selections
+apt-get install -y oracle-java8-installer
+apt-get install -y oracle-java8-set-default
+apt-get install -y maven
 ```
 
 
diff --git a/tests/ci_build/Dockerfile.ubuntu1404_cuda75_cudnn5 
b/tests/ci_build/Dockerfile.ubuntu1404_cuda75_cudnn5
index e9810af..88fd7ce 100644
--- a/tests/ci_build/Dockerfile.ubuntu1404_cuda75_cudnn5
+++ b/tests/ci_build/Dockerfile.ubuntu1404_cuda75_cudnn5
@@ -23,7 +23,13 @@ RUN cd /usr/src/gtest && cmake CMakeLists.txt && make && cp 
*.a /usr/lib
 RUN pip install nose cpplint 'pylint==1.4.4' 'astroid==1.3.6'
 
 # MAVEN
-RUN apt-get install -y maven default-jdk
+RUN apt-get install -y software-properties-common
+RUN add-apt-repository ppa:webupd8team/java -y
+RUN apt-get update
+RUN echo "oracle-java8-installer shared/accepted-oracle-license-v1-1 select 
true" | debconf-set-selections
+RUN apt-get install -y oracle-java8-installer
+RUN apt-get install -y oracle-java8-set-default
+RUN apt-get install -y maven
 
 # R
 RUN apt-get install -y software-properties-common r-base-core 
libcurl4-openssl-dev libssl-dev libxml2-dev
diff --git a/tests/ci_build/install/ubuntu_install_scala.sh 
b/tests/ci_build/install/ubuntu_install_scala.sh
index 712eff9..169ece0 100755
--- a/tests/ci_build/install/ubuntu_install_scala.sh
+++ b/tests/ci_build/install/ubuntu_install_scala.sh
@@ -19,5 +19,10 @@
 
 # install libraries for mxnet's scala package on ubuntu
 
-apt-get update && apt-get install -y \
-maven default-jdk
+apt-get install -y software-properties-common
+add-apt-repository -y ppa:webupd8team/java
+apt-get update
+echo "oracle-java8-installer shared/accepted-oracle-license-v1-1 select true" 
| debconf-set-selections
+apt-get install -y oracle-java8-installer
+apt-get install -y oracle-java8-set-default
+apt-get update && apt-get install -y maven

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

[GitHub] kevinthesun opened a new pull request #7636: Add script to build doc files for all versions

2017-08-27 Thread git 
kevinthesun opened a new pull request #7636: Add script to build doc files for 
all versions
URL: https://github.com/apache/incubator-mxnet/pull/7636
 
 
   Script to build all versions selected. Can be used for website recovering if 
apache jenkins automate building system generates unexpected result.
 

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 #7635: add fashion mnist and move mnists to s3

2017-08-27 Thread git 
piiswrong commented on a change in pull request #7635: add fashion mnist and 
move mnists to s3
URL: https://github.com/apache/incubator-mxnet/pull/7635#discussion_r135404465
 
 

 ##
 File path: python/mxnet/gluon/data/vision.py
 ##
 @@ -102,6 +102,56 @@ def _get_data(self):
 self._label = label
 
 
+class FashionMNIST(_DownloadedDataset):
+"""A dataset of Zalando's article images consisting of fashion products,
+a drop-in replacement of the original MNIST dataset from
+`https://github.com/zalandoresearch/fashion-mnist`_.
+
+Each sample is an image (in 3D NDArray) with shape (28, 28, 1).
+
+Parameters
+--
+root : str
+Path to temp folder for storing data.
+train : bool
+Whether to load the training or testing set.
+transform : function
+A user defined callback that transforms each instance. For example::
+
+transform=lambda data, label: (data.astype(np.float32)/255, label)
+"""
+def __init__(self, root='~/.mxnet/datasets/fashion-mnist', train=True,
+ transform=None):
+super(FashionMNIST, self).__init__(root, train, transform)
+
+def _get_data(self):
+if not os.path.isdir(self._root):
+os.makedirs(self._root)
+url = 
'https://apache-mxnet.s3.amazonaws.com/gluon/dataset/fashion-mnist/'
+if self._train:
+data_file = download(url+'train-images-idx3-ubyte.gz', self._root,
 
 Review comment:
   move url into MNIST._train_data_url/_train_label_url etc and inherit MNIST
 

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 #7633: fix tests

2017-08-27 Thread git 
piiswrong closed pull request #7633: fix tests
URL: https://github.com/apache/incubator-mxnet/pull/7633
 
 
   
 

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


With regards,
Apache Git Services

[incubator-mxnet] branch master updated: fix tests (#7633)

2017-08-27 Thread 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 9aa051c  fix tests (#7633)
9aa051c is described below

commit 9aa051c2e87d41b4f2a61fb62728ecdf364f8997
Author: Sheng Zha 
AuthorDate: Sun Aug 27 00:14:28 2017 -0700

fix tests (#7633)
---
 tests/python/gpu/test_operator_gpu.py |  4 ++--
 tests/python/unittest/test_loss.py| 18 ++
 2 files changed, 8 insertions(+), 14 deletions(-)

diff --git a/tests/python/gpu/test_operator_gpu.py 
b/tests/python/gpu/test_operator_gpu.py
index 11d146c..0c5771e 100644
--- a/tests/python/gpu/test_operator_gpu.py
+++ b/tests/python/gpu/test_operator_gpu.py
@@ -1346,11 +1346,11 @@ def test_sequence_reverse():
 
 
 def test_autograd_save_memory():
-x = mx.nd.zeros((128, 1024, 1024), ctx=mx.gpu(0))
+x = mx.nd.zeros((128, 512, 512), ctx=mx.gpu(0))
 x.attach_grad()
 
 with mx.autograd.record():
-for i in range(50):
+for i in range(200):
 x = x + 1
 x.wait_to_read()
 x.backward()
diff --git a/tests/python/unittest/test_loss.py 
b/tests/python/unittest/test_loss.py
index b864215..85875c6 100644
--- a/tests/python/unittest/test_loss.py
+++ b/tests/python/unittest/test_loss.py
@@ -63,7 +63,6 @@ def get_net(num_hidden):
 
 
 def test_ce_loss():
-mx.random.seed(1234)
 np.random.seed(1234)
 nclass = 10
 N = 20
@@ -83,7 +82,6 @@ def test_ce_loss():
 
 
 def test_bce_loss():
-mx.random.seed(1234)
 np.random.seed(1234)
 N = 20
 data = mx.random.uniform(-1, 1, shape=(N, 20))
@@ -111,7 +109,6 @@ def test_bce_equal_ce2():
 
 
 def test_kl_loss():
-mx.random.seed(1234)
 np.random.seed(1234)
 N = 20
 data = mx.random.uniform(-1, 1, shape=(N, 10))
@@ -129,12 +126,11 @@ def test_kl_loss():
 
 
 def test_l2_loss():
-mx.random.seed(1234)
 np.random.seed(1234)
 N = 20
 data = mx.random.uniform(-1, 1, shape=(N, 10))
 label = mx.random.uniform(-1, 1, shape=(N, 1))
-data_iter = mx.io.NDArrayIter(data, label, batch_size=10, 
label_name='label')
+data_iter = mx.io.NDArrayIter(data, label, batch_size=10, 
label_name='label', shuffle=True)
 output = get_net(1)
 l = mx.symbol.Variable('label')
 Loss = gluon.loss.L2Loss()
@@ -142,26 +138,25 @@ def test_l2_loss():
 loss = Loss(output, l)
 loss = mx.sym.make_loss(loss)
 mod = mx.mod.Module(loss, data_names=('data',), label_names=('label',))
-mod.fit(data_iter, num_epoch=200, optimizer_params={'learning_rate': 1.},
-eval_metric=mx.metric.Loss())
+mod.fit(data_iter, num_epoch=200, optimizer_params={'learning_rate': 0.1, 
'wd': 0.00045},
+initializer=mx.init.Xavier(magnitude=2), 
eval_metric=mx.metric.Loss())
 assert mod.score(data_iter, eval_metric=mx.metric.Loss())[0][1] < 0.05
 
 
 def test_l1_loss():
-mx.random.seed(1234)
 np.random.seed(1234)
 N = 20
 data = mx.random.uniform(-1, 1, shape=(N, 10))
 label = mx.random.uniform(-1, 1, shape=(N, 1))
-data_iter = mx.io.NDArrayIter(data, label, batch_size=10, 
label_name='label')
+data_iter = mx.io.NDArrayIter(data, label, batch_size=10, 
label_name='label', shuffle=True)
 output = get_net(1)
 l = mx.symbol.Variable('label')
 Loss = gluon.loss.L1Loss()
 loss = Loss(output, l)
 loss = mx.sym.make_loss(loss)
 mod = mx.mod.Module(loss, data_names=('data',), label_names=('label',))
-mod.fit(data_iter, num_epoch=200, optimizer_params={'learning_rate': 0.1},
-initializer=mx.init.Uniform(0.5), eval_metric=mx.metric.Loss())
+mod.fit(data_iter, num_epoch=200, optimizer_params={'learning_rate': 0.01},
+initializer=mx.init.Xavier(magnitude=3), 
eval_metric=mx.metric.Loss())
 assert mod.score(data_iter, eval_metric=mx.metric.Loss())[0][1] < 0.1
 
 
@@ -196,7 +191,6 @@ def test_ctc_loss():
 
 
 def test_sample_weight_loss():
-mx.random.seed(1234)
 np.random.seed(1234)
 nclass = 10
 N = 20

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

[GitHub] qingzhouzhen commented on issue #7632: pvanet infer_shape error. Arguments

2017-08-27 Thread git 
qingzhouzhen commented on issue #7632: pvanet infer_shape error. Arguments
URL: 
https://github.com/apache/incubator-mxnet/issues/7632#issuecomment-325180013
 
 
   I change the config.SCALES to 1056x640, It still went wrong with the same 
problem,  But I can print whole layers with mx.viz.print_summary()--It 
demonstrates my network has no concat problem, but I change to 600*1000, it 
went worong, is the system fit to any config.SCALES? 
   below layers when I  use 
   `mx.viz.print_summary(net, {"data":(1,3,1056,640),"gt_boxes": (1, 100, 5), 
"label": (1, 23760), "bbox_target": (1, 36, 66, 40), "bbox_weight": (1, 36, 66, 
40)})`, it seems no concat problem, but `mx.viz.print_summary(net, 
{"data":(1,3,600,1000),"gt_boxes": (1, 100, 5), "label": (1, 20646), 
"bbox_target": (1, 36, 37, 62), "bbox_weight": (1, 36, 37, 62)})` does not work
   
--
   `cls_prob(SoftmaxOutput) 1000
0   cls_score   

   custom34
   

   cls_prob_reshape(Reshape)   64x1000 
0   cls_prob
   

   bbox_pred(FullyConnected)   4000
16388096drop7   
   

   _minus75(_sub)  4000
0   bbox_pred   

   custom34
   

   bbox_loss_(smooth_l1)   4000
0   _minus75
   

   _mul75(_mul)4000
0   custom34

   bbox_loss_  
   

   bbox_loss(MakeLoss) 4000
0   _mul75  
   

   bbox_loss_reshape(Reshape)  64x4000 
0   bbox_loss   
   

   label_reshape(Reshape)  64  
0   custom34
   

   blockgrad9(BlockGrad)   64  
0   label_reshape   
   

   Total params: 143975542
   
`
   .
   .
   .
   too much to show 
   @reminisce 
 

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


With regards,
Apache Git Services

[GitHub] qingzhouzhen commented on issue #7632: pvanet infer_shape error. Arguments

2017-08-27 Thread git 
qingzhouzhen commented on issue #7632: pvanet infer_shape error. Arguments
URL: 
https://github.com/apache/incubator-mxnet/issues/7632#issuecomment-325180013
 
 
   I change the config.SCALES to 1056x640, It still went wrong with the same 
problem,  But I can print whole layers with mx.viz.print_summary(), but I 
change to 600*1000, it went worong, is the system fit to any config.SCALES? 
   below layers when I  use 
   `mx.viz.print_summary(net, {"data":(1,3,1056,640),"gt_boxes": (1, 100, 5), 
"label": (1, 23760), "bbox_target": (1, 36, 66, 40), "bbox_weight": (1, 36, 66, 
40)})`, it seems no concat problem, but `mx.viz.print_summary(net, 
{"data":(1,3,600,1000),"gt_boxes": (1, 100, 5), "label": (1, 20646), 
"bbox_target": (1, 36, 37, 62), "bbox_weight": (1, 36, 37, 62)})` does not work
   
--
   `cls_prob(SoftmaxOutput) 1000
0   cls_score   

   custom34
   

   cls_prob_reshape(Reshape)   64x1000 
0   cls_prob
   

   bbox_pred(FullyConnected)   4000
16388096drop7   
   

   _minus75(_sub)  4000
0   bbox_pred   

   custom34
   

   bbox_loss_(smooth_l1)   4000
0   _minus75
   

   _mul75(_mul)4000
0   custom34

   bbox_loss_  
   

   bbox_loss(MakeLoss) 4000
0   _mul75  
   

   bbox_loss_reshape(Reshape)  64x4000 
0   bbox_loss   
   

   label_reshape(Reshape)  64  
0   custom34
   

   blockgrad9(BlockGrad)   64  
0   label_reshape   
   

   Total params: 143975542
   
`
   .
   .
   .
   too much to show 
   @reminisce 
 

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


With regards,
Apache Git Services

[GitHub] szha opened a new pull request #7635: add fashion mnist and move mnists to s3

2017-08-27 Thread git 
szha opened a new pull request #7635: add fashion mnist and move mnists to s3
URL: https://github.com/apache/incubator-mxnet/pull/7635
 
 
   @mli 
 

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