fhieber commented on issue #6997: Generalized batch_dot / matrix multiplication
URL:
https://github.com/apache/incubator-mxnet/issues/6997#issuecomment-343069228
@nicklhy mxnet-0.12.0 added a `flatten` argument to FullyConnected. If set
to `False`, you can pass in n-dimensional input and
zhreshold commented on issue #8338: master branch cannot build on centos 7 with
cuda-8.0
URL:
https://github.com/apache/incubator-mxnet/issues/8338#issuecomment-343067371
@mseeger The guy reported it's solved after your PR, and I think it's
related to mshadow_op.h, that's why I am
mseeger commented on issue #8338: master branch cannot build on centos 7 with
cuda-8.0
URL:
https://github.com/apache/incubator-mxnet/issues/8338#issuecomment-343066287
Hello,
The issue was closed, because the person who raised it figured it out.
Please contact this person.
eric-haibin-lin commented on issue #8500: program crash when run sparse model
predict
URL:
https://github.com/apache/incubator-mxnet/issues/8500#issuecomment-343057807
Usually you want to increase the learning rate if batch size is increased.
What's the density of your model and
eric-haibin-lin closed issue #8491: Inconsistent behavior of sparse
elemwise_mul and dense elemwise_mul with nan
URL: https://github.com/apache/incubator-mxnet/issues/8491
This is an automated message from the Apache Git
eric-haibin-lin commented on a change in pull request #8539: expose group2ctx
to module
URL: https://github.com/apache/incubator-mxnet/pull/8539#discussion_r149873831
##
File path: tests/python/unittest/test_module.py
##
@@ -70,6 +70,63 @@ def test_module_input_grads():
eric-haibin-lin commented on a change in pull request #8539: expose group2ctx
to module
URL: https://github.com/apache/incubator-mxnet/pull/8539#discussion_r149873775
##
File path: python/mxnet/module/module.py
##
@@ -73,6 +75,11 @@ def __init__(self, symbol,
anxiaoxi45 commented on issue #5677: Constant Initializer only works on keys
ending with weight
URL:
https://github.com/apache/incubator-mxnet/issues/5677#issuecomment-343056608
@jmacglashanDo you have any new findings?
anxiaoxi45 commented on issue #5677: Constant Initializer only works on keys
ending with weight
URL:
https://github.com/apache/incubator-mxnet/issues/5677#issuecomment-343056448
I also met ... @tqchen @piiswrong @mli
This
eric-haibin-lin closed issue #8583: Improve custom cross-entropy operator
example
URL: https://github.com/apache/incubator-mxnet/issues/8583
This is an automated message from the Apache Git Service.
To respond to the
gnikh opened a new issue #8595: how to implement function like
numpy.random.choice using ndarray?
URL: https://github.com/apache/incubator-mxnet/issues/8595
I want to sample from some discrete elements, each of which with a given
probability, like the function of numpy:
lyblsgo commented on issue #8586: How to train Word-language-modeling-RNN on
multiple GPUs???
URL:
https://github.com/apache/incubator-mxnet/issues/8586#issuecomment-343025596
@eric-haibin-lin Can you give me some advice?
tz-hmc opened a new issue #8594: How do I extract layers and their
"functionality" from pre-trained network?
URL: https://github.com/apache/incubator-mxnet/issues/8594
## Description:
I need to make a Siamese network for image similarity, and to do so I'm
hoping to use 2 copies of
nicklhy commented on issue #6997: Generalized batch_dot / matrix multiplication
URL:
https://github.com/apache/incubator-mxnet/issues/6997#issuecomment-343017596
@fhieber , have you figured out a solution about this feature now?
piiswrong commented on a change in pull request #8294: NCCL integration
URL: https://github.com/apache/incubator-mxnet/pull/8294#discussion_r149841891
##
File path: python/mxnet/model.py
##
@@ -104,15 +105,18 @@ def _initialize_kvstore(kvstore, param_arrays,
arg_params,
reminisce commented on a change in pull request #8566: optimize broadcast
URL: https://github.com/apache/incubator-mxnet/pull/8566#discussion_r149839401
##
File path: src/operator/mxnet_op.h
##
@@ -345,6 +394,13 @@ __global__ void mxnet_generic_kernel(int N, Args... args)
szha closed issue #5285: train_cifar10.py not running
URL: https://github.com/apache/incubator-mxnet/issues/5285
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and
szha commented on issue #5285: train_cifar10.py not running
URL:
https://github.com/apache/incubator-mxnet/issues/5285#issuecomment-343007330
This issue is closed due to lack of activity in the last 90 days. Feel free
to ping me to reopen if this is still an active issue. Thanks!
Also,
jdchoi77 opened a new issue #8593: Runtime errors during forward in a custom
gluon.Block
URL: https://github.com/apache/incubator-mxnet/issues/8593
I created a dummy `Block`, which takes a 2D array, performs a 2D
convolution, and feeds the convoluted output to a fully connected layer :
szha commented on issue #7957: add densenet
URL: https://github.com/apache/incubator-mxnet/pull/7957#issuecomment-34201
@piiswrong shall we merge this?
This is an automated message from the Apache Git Service.
To respond
yajiedesign commented on a change in pull request #8572: use first class cuda
with cmake 3.9 and cuda9.0 support
URL: https://github.com/apache/incubator-mxnet/pull/8572#discussion_r149830346
##
File path: cmake/FirstCuda.cmake
##
@@ -0,0 +1,233 @@
+# Licensed to the
yajiedesign commented on a change in pull request #8572: use first class cuda
with cmake 3.9 and cuda9.0 support
URL: https://github.com/apache/incubator-mxnet/pull/8572#discussion_r149829711
##
File path: CMakeLists.txt
##
@@ -128,14 +152,20 @@ endif()
yajiedesign commented on a change in pull request #8572: use first class cuda
with cmake 3.9 and cuda9.0 support
URL: https://github.com/apache/incubator-mxnet/pull/8572#discussion_r149829711
##
File path: CMakeLists.txt
##
@@ -128,14 +152,20 @@ endif()
tz-hmc commented on issue #8591: How do I make a siamese network with
pretrained models (esp. keeping the weights the same?)
URL:
https://github.com/apache/incubator-mxnet/issues/8591#issuecomment-342970864
Wow, I didn't know that API existed. I had a lot of trouble trying to make
it
piiswrong commented on a change in pull request #8566: optimize broadcast
URL: https://github.com/apache/incubator-mxnet/pull/8566#discussion_r149823235
##
File path: src/operator/mxnet_op.h
##
@@ -345,6 +394,13 @@ __global__ void mxnet_generic_kernel(int N, Args... args)
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 6da0472 reimplement cross-entropy
piiswrong closed pull request #8588: reimplement cross-entropy operator in
recommender example with NDArray
URL: https://github.com/apache/incubator-mxnet/pull/8588
This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the
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 8dc09ec gluon with multiple data type
piiswrong closed pull request #8522: gluon with multiple data type
URL: https://github.com/apache/incubator-mxnet/pull/8522
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
anirudh2290 commented on issue #8592: Simple bind doesnt infer the provided
stype for non default stypes
URL:
https://github.com/apache/incubator-mxnet/issues/8592#issuecomment-342988942
@eric-haibin-lin
This is an
anirudh2290 opened a new issue #8592: Simple bind doesnt infer the provided
stype for non default stypes
URL: https://github.com/apache/incubator-mxnet/issues/8592
## Description
Simple bind doesn't infer the provided stype for non default stypes
## Environment info (Required)
zhreshold commented on issue #8338: master branch cannot build on centos 7 with
cuda-8.0
URL:
https://github.com/apache/incubator-mxnet/issues/8338#issuecomment-342985714
@mseeger Have you figured out what is the cause of this problem?
I encountered exactly same error during CI
piiswrong closed pull request #8558: slice operator supporting arbitrary values
of step
URL: https://github.com/apache/incubator-mxnet/pull/8558
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 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 70b68b1 slice operator supporting
tz-hmc commented on issue #8591: How do I make a siamese network with
pretrained models (esp. keeping the weights the same?)
URL:
https://github.com/apache/incubator-mxnet/issues/8591#issuecomment-342970864
Wow, I didn't know that API existed. I had a lot of trouble trying to make
it
tz-hmc commented on issue #8591: How do I make a siamese network with
pretrained models (esp. keeping the weights the same?)
URL:
https://github.com/apache/incubator-mxnet/issues/8591#issuecomment-342970864
I had a lot of trouble trying to make it work with the module API. But the
Gluon
edmBernard commented on issue #8591: How do I make a siamese network with
pretrained models (esp. keeping the weights the same?)
URL:
https://github.com/apache/incubator-mxnet/issues/8591#issuecomment-342959619
I got a similar problem and achieve to got several solutions:
zhreshold commented on issue #8589: [Discussion] Data Pipeline Intermediate
Representation in MXNet/NNVM
URL:
https://github.com/apache/incubator-mxnet/issues/8589#issuecomment-342957255
Data preprocessing operator is under dev, until then it's hard to pack data
pipeline into graph.
zhreshold commented on issue #8546: What's the equivalent of np.where()?
URL:
https://github.com/apache/incubator-mxnet/issues/8546#issuecomment-342956240
Not sure how you would do with the cls_scores, if you can continue use
masks, then nd.where is good, if you would like to strip out
jingpengw commented on issue #6872: HDF5 DataIter Python
URL:
https://github.com/apache/incubator-mxnet/issues/6872#issuecomment-342935141
@xiaogehh I think the provide_data function was required and used by mxnet
dataprovider. Whenever you create a new data iter, you need to define
eric-haibin-lin closed issue #8590: the error:mshadow::cuda::AddTakeGrad' :
ambiguous call to overloaded function indexing_op.h
URL: https://github.com/apache/incubator-mxnet/issues/8590
This is an automated message from
rahul003 closed pull request #8507: Passing user specified environment
variables when using launch.py
URL: https://github.com/apache/incubator-mxnet/pull/8507
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
zhreshold commented on issue #8582: Yolo2 operator
URL: https://github.com/apache/incubator-mxnet/pull/8582#issuecomment-342920908
@sbodenstein Yes, I agree with your comments. I am working on object
detection operators decomposed and implemented using gluon blocks/nnvm
operators that is
sbodenstein commented on issue #8582: Yolo2 operator
URL: https://github.com/apache/incubator-mxnet/pull/8582#issuecomment-342901844
One issue for us: this layer is an 'output layer' and hence doesn't support
out grads into the layer. This is an unfortunate restriction, as it rules out
loweew commented on issue #7933: CoreML conversion with finetuned model
URL:
https://github.com/apache/incubator-mxnet/issues/7933#issuecomment-342895535
any update here? do you need me to provide a model for you to move forward?
cjolivier01 commented on a change in pull request #8572: use first class cuda
with cmake 3.9 and cuda9.0 support
URL: https://github.com/apache/incubator-mxnet/pull/8572#discussion_r149741904
##
File path: cmake/FirstCuda.cmake
##
@@ -0,0 +1,233 @@
+# Licensed to the
cjolivier01 commented on a change in pull request #8572: use first class cuda
with cmake 3.9 and cuda9.0 support
URL: https://github.com/apache/incubator-mxnet/pull/8572#discussion_r149741331
##
File path: CMakeLists.txt
##
@@ -128,14 +152,20 @@ endif()
ShownX commented on issue #8193: add center loss
URL: https://github.com/apache/incubator-mxnet/pull/8193#issuecomment-342884841
changes made here: https://github.com/ShownX/mxnet-center-loss. Closed.
This is an automated
ShownX closed pull request #8193: add center loss
URL: https://github.com/apache/incubator-mxnet/pull/8193
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
aaronmarkham commented on a change in pull request #8578: Add development
instructions and update installation instructions
URL: https://github.com/apache/incubator-mxnet/pull/8578#discussion_r149700972
##
File path: DEVELOPMENT.md
##
@@ -0,0 +1,69 @@
+# Developing MXNet
aaronmarkham commented on a change in pull request #8578: Add development
instructions and update installation instructions
URL: https://github.com/apache/incubator-mxnet/pull/8578#discussion_r149698522
##
File path: DEVELOPMENT.md
##
@@ -0,0 +1,69 @@
+# Developing MXNet
aaronmarkham commented on a change in pull request #8578: Add development
instructions and update installation instructions
URL: https://github.com/apache/incubator-mxnet/pull/8578#discussion_r149697872
##
File path: DEVELOPMENT.md
##
@@ -0,0 +1,69 @@
+# Developing MXNet
aaronmarkham commented on a change in pull request #8578: Add development
instructions and update installation instructions
URL: https://github.com/apache/incubator-mxnet/pull/8578#discussion_r149698741
##
File path: DEVELOPMENT.md
##
@@ -0,0 +1,69 @@
+# Developing MXNet
aaronmarkham commented on a change in pull request #8578: Add development
instructions and update installation instructions
URL: https://github.com/apache/incubator-mxnet/pull/8578#discussion_r149699813
##
File path: DEVELOPMENT.md
##
@@ -0,0 +1,69 @@
+# Developing MXNet
aaronmarkham commented on a change in pull request #8578: Add development
instructions and update installation instructions
URL: https://github.com/apache/incubator-mxnet/pull/8578#discussion_r149696564
##
File path: docs/install/index.md
##
@@ -65,36 +65,37 @@ Indicate
szha commented on issue #5837: Error caused by two different dependencies in
protobuf version when using both distributed mode and tensorboard
URL:
https://github.com/apache/incubator-mxnet/issues/5837#issuecomment-342802907
This issue is closed due to lack of activity in the last 90
szha commented on issue #6999: bgr or rgb input to vgg16?
URL:
https://github.com/apache/incubator-mxnet/issues/6999#issuecomment-342802911
This issue is closed due to lack of activity in the last 90 days. Feel free
to ping me to reopen if this is still an active issue. Thanks!
Also,
szha closed issue #6999: bgr or rgb input to vgg16?
URL: https://github.com/apache/incubator-mxnet/issues/6999
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use
szha closed issue #5837: Error caused by two different dependencies in protobuf
version when using both distributed mode and tensorboard
URL: https://github.com/apache/incubator-mxnet/issues/5837
This is an automated
ZiyueHuang commented on a change in pull request #8259: check_format of sparse
ndrray
URL: https://github.com/apache/incubator-mxnet/pull/8259#discussion_r149650525
##
File path: src/common/utils.h
##
@@ -43,9 +43,170 @@
#include
#include
+#include
larroy commented on issue #8554: Disable long running tests
URL: https://github.com/apache/incubator-mxnet/pull/8554#issuecomment-342789057
I think this is a good approach. I would like to see this merged. >30s or
too high of resource consumption for a unit test is not acceptable.
dmas-at-wiris commented on issue #7551: How to get the symbol of gluon's RNN
model by hybridizing
URL:
https://github.com/apache/incubator-mxnet/issues/7551#issuecomment-342776563
@piiswrong Any ETA for this feature?
This
KellenSunderland commented on issue #8554: Disable long running tests
URL: https://github.com/apache/incubator-mxnet/pull/8554#issuecomment-342756833
@piiswrong I think having a multi-stage approach to speeding up tests is
going to be the best way to make progress here. This first stage
KellenSunderland commented on a change in pull request #8582: Yolo2 operator
URL: https://github.com/apache/incubator-mxnet/pull/8582#discussion_r149613155
##
File path: src/operator/contrib/yolo_output-inl.h
##
@@ -0,0 +1,702 @@
+/*
+ * Licensed to the Apache Software
KellenSunderland commented on issue #8554: Disable long running tests
URL: https://github.com/apache/incubator-mxnet/pull/8554#issuecomment-342756833
@piiswrong I think having a multi-stage approach to speeding up tests is
going to be the best way to make progress here. This first stage
KellenSunderland commented on issue #8554: Disable long running tests
URL: https://github.com/apache/incubator-mxnet/pull/8554#issuecomment-342756833
@piiswrong I think having a multi-stage approach to speeding up tests is
going to be the best way to make progress here. This first stage
KellenSunderland commented on issue #8554: Disable long running tests
URL: https://github.com/apache/incubator-mxnet/pull/8554#issuecomment-342756833
@piiswrong I think having a multi-stage approach to speeding up tests is
going to be the best way to make progress here. This first stage
Jing-Luo commented on issue #1356: override global initialization method in
layer configuration
URL:
https://github.com/apache/incubator-mxnet/issues/1356#issuecomment-342756881
```
init = mx.initializer.Mixed(['bias', '.*'], [mx.init.Zero(),
mx.init.Xavier(magnitude=2.24)])
Jing-Luo commented on issue #1356: override global initialization method in
layer configuration
URL:
https://github.com/apache/incubator-mxnet/issues/1356#issuecomment-342756881
```
init = mx.initializer.Mixed(['bias', '.*'], [mx.init.Zero(),
mx.init.Xavier(magnitude=2.24)])
KellenSunderland commented on issue #8554: Disable long running tests
URL: https://github.com/apache/incubator-mxnet/pull/8554#issuecomment-342756833
@piiswrong I think having a multi-stage approach to speeding up tests is
going to be the best way to make progress here. This first stage
994088895 commented on issue #8590: the error:mshadow::cuda::AddTakeGrad' :
ambiguous call to overloaded function indexing_op.h
URL:
https://github.com/apache/incubator-mxnet/issues/8590#issuecomment-342756960
Thank you, I have resoved it. It seems that I wrongly used the 'git clone
KellenSunderland commented on issue #8554: Disable long running tests
URL: https://github.com/apache/incubator-mxnet/pull/8554#issuecomment-342756833
@piiswrong I think having a multi-stage approach to speeding up tests is
going to be the best way to make progress here. This first stage
qxde01 commented on issue #8234: A bug in VGG_FC_ILSVRC_16_layers-symbol.json
URL:
https://github.com/apache/incubator-mxnet/issues/8234#issuecomment-342753475
I catched an error of VGG_FC_ILSVRC_16_layers?
`Traceback (most recent call last):`
`File "", line 1, in `
`File
qxde01 commented on issue #8234: A bug in VGG_FC_ILSVRC_16_layers-symbol.json
URL:
https://github.com/apache/incubator-mxnet/issues/8234#issuecomment-342753475
I also catched this error of VGG_FC_ILSVRC_16_layers?
`Traceback (most recent call last):`
`File "", line 1, in `
qxde01 commented on issue #8234: A bug in VGG_FC_ILSVRC_16_layers-symbol.json
URL:
https://github.com/apache/incubator-mxnet/issues/8234#issuecomment-342753475
I also catched this error of VGG_FC_ILSVRC_16_layers?
`Traceback (most recent call last):`
`File "", line 1, in `
asmushetzel commented on a change in pull request #8577: support for lapack
functions with mkl
URL: https://github.com/apache/incubator-mxnet/pull/8577#discussion_r149604263
##
File path: src/operator/c_lapack_api.h
##
@@ -160,7 +162,73 @@ inline void flip
asmushetzel commented on a change in pull request #8577: support for lapack
functions with mkl
URL: https://github.com/apache/incubator-mxnet/pull/8577#discussion_r149603534
##
File path: src/operator/c_lapack_api.h
##
@@ -271,6 +339,9 @@ inline void flip(int
asmushetzel commented on a change in pull request #8577: support for lapack
functions with mkl
URL: https://github.com/apache/incubator-mxnet/pull/8577#discussion_r149603492
##
File path: src/operator/c_lapack_api.h
##
@@ -160,7 +162,73 @@ inline void flip
asmushetzel commented on a change in pull request #8577: support for lapack
functions with mkl
URL: https://github.com/apache/incubator-mxnet/pull/8577#discussion_r149603408
##
File path: make/config.mk
##
@@ -105,6 +105,12 @@ USE_LAPACK = 1
# path to lapack library in
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-342745203
it did, but I just went and removed all mxnet files and reinstalled via pip.
Seems to work now. Sorry :)
SumNeuron closed issue #7900: Request: finish python gpu enabled guide for
install
URL: https://github.com/apache/incubator-mxnet/issues/7900
This is an automated message from the Apache Git Service.
To respond to the
szha commented on issue #7900: Request: finish python gpu enabled guide for
install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-342744184
Does it show any other error, such as "libxxx not found"?
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-342743473
If I remove the director mxnet loads, but then it says things like
```
import mxnet as mx
from mxnet
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-342742633
I dont think so. How would I check for this?
This is
szha commented on issue #7900: Request: finish python gpu enabled guide for
install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-342742311
Sorry for missing on this. From the error it looks like your python is
trying to load from `/home/sumner/mxnet/`, which
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-342741096
@szha this error is still presistent for `pip3 install --pre mxnet-cu90`
SumNeuron opened a new issue #7900: Request: finish python gpu enabled guide
for install
URL: https://github.com/apache/incubator-mxnet/issues/7900
Please complete the python GPU enabled install for MacOS.
This guide is complete / following it to complete results in errors.
87 matches
Mail list logo