chowkamlee81 commented on issue #7907: Dense upsamling operation rather than
deconvolution implementation
URL:
https://github.com/apache/incubator-mxnet/issues/7907#issuecomment-330437118
Hai szha,
I tried according to you and i removed crop function as you suggested. After
that
chowkamlee81 commented on issue #7717: Subpixel convolution(state of art)
implementation rather than using Deconvolution.
URL:
https://github.com/apache/incubator-mxnet/issues/7717#issuecomment-330436790
Dear Eldercrow,
I tried according to you and i removed crop function as you
cjolivier01 commented on issue #7935: Zeroes CSR still needs a valid
row_pointer array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#issuecomment-330434805
I don't see any burden on memory allocation greater than normal
NDArray.zeros() allocation. It'll be less, actually,
reminisce commented on issue #7935: Zeroes CSR still needs a valid row_pointer
array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#issuecomment-330423935
I agree that an all zero initialized `indptr` in a zero csr sounds more
correct from the definition point of view of csr
mbaijal commented on issue #7941: DO NOT MERGE: Just a test of adding
properties to JenkinsFile
URL: https://github.com/apache/incubator-mxnet/pull/7941#issuecomment-330403575
Build Now
This is an automated message from the
mbaijal opened a new pull request #7941: DO NOT MERGE: Just a test of adding
properties to JenkinsFile
URL: https://github.com/apache/incubator-mxnet/pull/7941
Please DO NOT MERGE!!
This is an automated message from the
cjolivier01 commented on issue #7935: Zeroes CSR still needs a valid
row_pointer array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#issuecomment-330400617
While an unintialized RSP matrix is a valid all-zero matrix, this is not the
case for CSR. Thus RSP returns a valid
cjolivier01 commented on issue #7935: Zeroes CSR still needs a valid
row_pointer array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#issuecomment-330398788
I respectfully disagree. Maybe something like empty() should return
uninitialized, but returning an Invald CSR array from
thirdwing commented on issue #7476: R-package RNN refactor
URL: https://github.com/apache/incubator-mxnet/pull/7476#issuecomment-330396699
@jeremiedb Can you update with master branch? Let's see the CI test.
This is an
reminisce commented on issue #7682: Fix shape inference bug
URL: https://github.com/apache/incubator-mxnet/pull/7682#issuecomment-330377972
@tqchen I will do that. Thanks.
This is an automated message from the Apache Git
szha commented on issue #7938: instance norm and reflection padding
URL: https://github.com/apache/incubator-mxnet/pull/7938#issuecomment-33038
Could you fix lint? The errors can be obtained by running `make pylint`
This
tqchen commented on issue #7682: Fix shape inference bug
URL: https://github.com/apache/incubator-mxnet/pull/7682#issuecomment-330377101
@reminisce Can you also backport this to nnvm?
This is an automated message from the
eric-haibin-lin commented on a change in pull request #7935: Zeroes CSR still
needs a valid row_pointer array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#discussion_r139555468
##
File path: src/operator/tensor/init_op.h
##
@@ -197,14 +197,21 @@ void
cjolivier01 commented on a change in pull request #7935: Zeroes CSR still needs
a valid row_pointer array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#discussion_r139555019
##
File path: src/operator/tensor/init_op.h
##
@@ -197,14 +197,21 @@ void
reminisce commented on issue #7935: Zeroes CSR still needs a valid row_pointer
array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#issuecomment-330371832
I think it's okay for operators to work on a zero csr with empty `indptr` as
long as the csr is checked if zero before
cjolivier01 commented on a change in pull request #7935: Zeroes CSR still needs
a valid row_pointer array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#discussion_r139554395
##
File path: src/operator/tensor/init_op.h
##
@@ -197,14 +197,21 @@ void
eric-haibin-lin commented on a change in pull request #7935: Zeroes CSR still
needs a valid row_pointer array.
URL: https://github.com/apache/incubator-mxnet/pull/7935#discussion_r139547763
##
File path: src/operator/tensor/init_op.h
##
@@ -197,14 +197,21 @@ void
lichen11 commented on issue #6943: How to read rec files in R
URL:
https://github.com/apache/incubator-mxnet/issues/6943#issuecomment-330366761
Hi, I have a follow-up question:
dataiter <- mx.io.ImageRecordIter(
path.imgrec = "./data/cifar/train.rec",
path.imglist=
lichen11 commented on issue #6943: How to read rec files in R
URL:
https://github.com/apache/incubator-mxnet/issues/6943#issuecomment-330366761
Hi, I have a follow-up question:
dataiter <- mx.io.ImageRecordIter(
path.imgrec = "./data/cifar/train.rec",
path.imglist=
larroy commented on issue #7852: Trouble installing MXNet on Raspberry Pi 3
URL:
https://github.com/apache/incubator-mxnet/issues/7852#issuecomment-330361445
you should install as root or using a virtualenv
This is an
ykim362 commented on a change in pull request #7931: MKL-DNN integration:
request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139533086
##
File path: prepare_mkl.sh
##
@@ -115,7 +115,7 @@ if [ -z $MKLROOT ]; then
fi
# Check what
ykim362 commented on a change in pull request #7931: MKL-DNN integration:
request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139530795
##
File path: src/operator/mkl/mkl_conv-common-inl.h
##
@@ -0,0 +1,82 @@
ykim362 commented on a change in pull request #7931: MKL-DNN integration:
request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139526050
##
File path: src/operator/mkl/mkl_conv-common-inl.h
##
@@ -0,0 +1,82 @@
ykim362 commented on issue #7931: MKL-DNN integration: request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#issuecomment-330340158
@piiswrong Yes, MKL-DNN build passes 'test_operator.py' unit test same as
MKLML build. So, we are doing several experiments.
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330339788
Trying to import ndarray to do the test from the python subdir in the local
mxnet results in a error from ctypes...
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330336051
which is the right version of python :P
This is an
szha commented on issue #7900: Request: finish python gpu enabled guide for
install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330335895
setting your path to use the right copy of python should do. PATH setting
should be in your .bashrc/.zshrc
zhanghang1989 commented on issue #7570: Gluon InstanceNorm and
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#issuecomment-330335943
Got messed up ... Creating a new PR
This is an automated
zhanghang1989 closed pull request #7570: Gluon InstanceNorm and
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570
This is an automated message from the Apache Git Service.
To respond to the
szha commented on issue #7931: MKL-DNN integration: request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#issuecomment-330334021
@ykim362 could you add a job for testing mxnet with mkldnn in Jenkinsfile?
Otherwise the tests don't reflect the change.
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139519618
##
File path: src/operator/concat.cc
##
@@ -50,6 +55,18 @@ Operator* CreateOp(ConcatParam param, int
piiswrong commented on issue #7931: MKL-DNN integration: request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#issuecomment-330333421
Does consistency tests pass for mkldnn operators? If it passes it should
converge
piiswrong commented on issue #7931: MKL-DNN integration: request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#issuecomment-330333421
Does consistency tests pass for mkldnn operators?
This is an
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330333210
Good catch.
Dumb question: how do I take the master mxnet installed above and replace
the other mxnet...
szha commented on issue #7900: Request: finish python gpu enabled guide for
install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330332132
You seem to have two copies of python in your system, one in
`/usr/local/lib/python3.6/` and the other in
ykim362 commented on a change in pull request #7931: MKL-DNN integration:
request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139516810
##
File path: src/operator/concat.cc
##
@@ -50,6 +55,18 @@ Operator* CreateOp(ConcatParam param,
ykim362 commented on a change in pull request #7931: MKL-DNN integration:
request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139516729
##
File path: src/operator/activation.cc
##
@@ -29,12 +29,30 @@
#include "./mkl/mkl_memory-inl.h"
ykim362 commented on a change in pull request #7931: MKL-DNN integration:
request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139516496
##
File path: src/operator/activation.cc
##
@@ -29,12 +29,30 @@
#include "./mkl/mkl_memory-inl.h"
sbodenstein commented on issue #7931: MKL-DNN integration: request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#issuecomment-330329680
@szha: I asked the same thing, this was the response:
https://github.com/01org/mkl-dnn/issues/10#issuecomment-323730811
szha commented on a change in pull request #7903: Refactor AdaGrad optimizer to
support sparse tensors
URL: https://github.com/apache/incubator-mxnet/pull/7903#discussion_r139514542
##
File path: python/mxnet/optimizer.py
##
@@ -665,26 +667,46 @@ class AdaGrad(Optimizer):
szha commented on issue #7931: MKL-DNN integration: request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#issuecomment-330324846
General comment: I didn't see mac being supported on mkl-dnn page. It only
says the software was validated on RHEL7. What does this change
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139511137
##
File path: src/operator/mkl/mkl_memory.h
##
@@ -112,6 +117,10 @@ struct MKLMemHolder {
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139510894
##
File path: src/operator/mkl/mkl_conv-common-inl.h
##
@@ -0,0 +1,82 @@
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139510706
##
File path: src/operator/lrn.cc
##
@@ -40,6 +46,9 @@ Operator* CreateOp(LRNParam param, int dtype) {
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139510469
##
File path: src/operator/concat.cc
##
@@ -50,6 +55,18 @@ Operator* CreateOp(ConcatParam param, int
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139509832
##
File path: prepare_mkl.sh
##
@@ -115,7 +115,7 @@ if [ -z $MKLROOT ]; then
fi
# Check what MKL
szha commented on issue #7570: Gluon InstanceNorm and ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#issuecomment-330321800
I suppose this means you passed. Congrats :)
Could you do a rebase onto the latest master to let the tests run?
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139508448
##
File path: src/operator/activation.cc
##
@@ -29,12 +29,30 @@
#include "./mkl/mkl_memory-inl.h"
ykim362 commented on a change in pull request #7931: MKL-DNN integration:
request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139507607
##
File path: prepare_mkl.sh
##
@@ -115,7 +115,7 @@ if [ -z $MKLROOT ]; then
fi
# Check what
ykim362 commented on issue #7931: MKL-DNN integration: request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#issuecomment-330319229
@piiswrong MKL-DNN doesn't converge with Resnet. It converges with mnist. I
am looking into it to find a root cause. MKLML works fine
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139506527
##
File path: prepare_mkl.sh
##
@@ -115,7 +115,7 @@ if [ -z $MKLROOT ]; then
fi
# Check what MKL
szha commented on a change in pull request #7931: MKL-DNN integration: request
for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139506331
##
File path: prepare_mkl.sh
##
@@ -115,7 +115,7 @@ if [ -z $MKLROOT ]; then
fi
# Check what MKL
gautamkmr opened a new pull request #7937: Increase the tolerance
URL: https://github.com/apache/incubator-mxnet/pull/7937
@piiswrong @cjolivier01
This is an automated message from the Apache Git Service.
To respond to the
gautamkmr commented on issue #7926: Increase the tolerance
URL: https://github.com/apache/incubator-mxnet/pull/7926#issuecomment-330315224
Whops wrong test fix.
This is an automated message from the Apache Git Service.
To
gautamkmr commented on issue #7926: Increase the tolerance
URL: https://github.com/apache/incubator-mxnet/pull/7926#issuecomment-330313865
The test has been failing continuously here is another instance today
gautamkmr commented on issue #7926: Increase the tolerance
URL: https://github.com/apache/incubator-mxnet/pull/7926#issuecomment-330313317
@cjolivier01 I saw the failure since last two weeks, haven't noticed
before.
piiswrong commented on issue #7931: MKL-DNN integration: request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#issuecomment-330308578
what's the issue with resnet convergence?
This is an automated
zhanghang1989 commented on issue #7570: Gluon InstanceNorm and
ReflectancePadding
URL: https://github.com/apache/incubator-mxnet/pull/7570#issuecomment-330307254
Hey folks, I was busy with my thesis defense last week. Any further
feedbacks for this PR?
piiswrong commented on a change in pull request #7931: MKL-DNN integration:
request for reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931#discussion_r139495738
##
File path: src/operator/mkl/mkldnn_memory.cc
##
@@ -0,0 +1,285 @@
szha opened a new pull request #7936: simplify CTC forward after namespace
reorganize
URL: https://github.com/apache/incubator-mxnet/pull/7936
This is an automated message from the Apache Git Service.
To respond to the
cjolivier01 commented on a change in pull request #7903: Refactor AdaGrad
optimizer to support sparse tensors
URL: https://github.com/apache/incubator-mxnet/pull/7903#discussion_r139483314
##
File path: python/mxnet/optimizer.py
##
@@ -665,26 +667,46 @@ class
cjolivier01 commented on a change in pull request #7903: Refactor AdaGrad
optimizer to support sparse tensors
URL: https://github.com/apache/incubator-mxnet/pull/7903#discussion_r139483293
##
File path: python/mxnet/optimizer.py
##
@@ -665,26 +667,46 @@ class
FrancisTse8 commented on issue #7852: Trouble installing MXNet on Raspberry Pi 3
URL:
https://github.com/apache/incubator-mxnet/issues/7852#issuecomment-330285714
When I was running the make file, I was able to generate a libmxnet.so file.
Is this the .so file you are referring to?
cjolivier01 commented on issue #7920: csr binary operator bug
URL:
https://github.com/apache/incubator-mxnet/issues/7920#issuecomment-330284076
**Looks like problem was introduced here**
https://github.com/apache/incubator-mxnet/commit/0b1363116c84dcefa751a925749b2da04c3f2614
cjolivier01 opened a new pull request #7935: Zeroes CSR still needs a valid
row_pointer array.
URL: https://github.com/apache/incubator-mxnet/pull/7935
Fix for: https://github.com/apache/incubator-mxnet/issues/7920
**Looks like problem was introduced here**
x1year opened a new issue #7934: Bug of group2ctx? Wrong device placement?
URL: https://github.com/apache/incubator-mxnet/issues/7934
For the following code:
x = mx.symbol.MyOp()
exe = x.bind(mx.gpu(), {}, group2ctx={"a": mx.cpu(), "b": mx.gpu()})
exe.forward()
where
reminisce commented on a change in pull request #7911: More sparse related docs
URL: https://github.com/apache/incubator-mxnet/pull/7911#discussion_r139467525
##
File path: python/mxnet/ndarray/sparse.py
##
@@ -88,6 +88,8 @@ def _new_alloc_handle(stype, shape, ctx,
cjolivier01 commented on a change in pull request #7903: Refactor AdaGrad
optimizer to support sparse tensors
URL: https://github.com/apache/incubator-mxnet/pull/7903#discussion_r139461575
##
File path: python/mxnet/optimizer.py
##
@@ -665,26 +667,46 @@ class
cjolivier01 commented on a change in pull request #7903: Refactor AdaGrad
optimizer to support sparse tensors
URL: https://github.com/apache/incubator-mxnet/pull/7903#discussion_r139461575
##
File path: python/mxnet/optimizer.py
##
@@ -665,26 +667,46 @@ class
loweew opened a new issue #7933: CoreML conversion with finetuned model
URL: https://github.com/apache/incubator-mxnet/issues/7933
I have successfully converted the squeezenet and resnet50 models from the
examples to CoreML using mxnet-to-coreml. However, when converting a model
after
jiarenyf commented on issue #7925: Resize image to fixed size
URL:
https://github.com/apache/incubator-mxnet/issues/7925#issuecomment-330253920
Maybe this can help you:
"https://github.com/jiarenyf/mxWrapper/blob/02fd9b0fcd37f7224648efad651a6f83a1f06d78/mxHelper/mxData.py#L93;
.
cjolivier01 commented on issue #7920: csr binary operator bug
URL:
https://github.com/apache/incubator-mxnet/issues/7920#issuecomment-330251762
Taking a look...
This is an automated message from the Apache Git Service.
To
jiarenyf commented on issue #7710: CTC ERROR WITH CUDA ILLEGAL MEMORY ACCESS
ERROR
URL:
https://github.com/apache/incubator-mxnet/issues/7710#issuecomment-330248910
@szha
I found another reason of that problem.
As in
ZiyueHuang commented on issue #7925: Resize image to fixed size
URL:
https://github.com/apache/incubator-mxnet/issues/7925#issuecomment-330240283
You can use cv2.imresize for convenience. Methods in mx.image use mx.ndarray
internally for effiency. You can refer to mx.image.ImageIter in
Prasad9 commented on issue #7925: Resize image to fixed size
URL:
https://github.com/apache/incubator-mxnet/issues/7925#issuecomment-330237849
@ZiyueHuang Thanks for the method. That is what I had wanted. But now, I am
unable to understand how to use this `Augmenter` class. Because, there
ZiyueHuang commented on issue #7925: Resize image to fixed size
URL:
https://github.com/apache/incubator-mxnet/issues/7925#issuecomment-330216607
`mx.image.ForceResizeAug`. It uses an internal method
`mx.nd._internal._cvimresize`.
lijuan123 commented on issue #6474: Fine-tune the mxnet ssd get
mismatchfrom.shape() error
URL:
https://github.com/apache/incubator-mxnet/issues/6474#issuecomment-330204414
@liumusicforever thank you!
This is an automated
liumusicforever commented on issue #6474: Fine-tune the mxnet ssd get
mismatchfrom.shape() error
URL:
https://github.com/apache/incubator-mxnet/issues/6474#issuecomment-330188579
I solve from making sure the `num_class` is equal with calling by symbol.py
(importlib)"" and loading from
liumusicforever commented on issue #6474: Fine-tune the mxnet ssd get
mismatchfrom.shape() error
URL:
https://github.com/apache/incubator-mxnet/issues/6474#issuecomment-330201273
sorry , I make a mistake , I mean **numbers of class** not **shape of data**
above.
liumusicforever commented on issue #6474: Fine-tune the mxnet ssd get
mismatchfrom.shape() error
URL:
https://github.com/apache/incubator-mxnet/issues/6474#issuecomment-330200756
Did your classes number of pretrained model is same as classes number of
symbol ?
lijuan123 commented on issue #6474: Fine-tune the mxnet ssd get
mismatchfrom.shape() error
URL:
https://github.com/apache/incubator-mxnet/issues/6474#issuecomment-330199777
@liumusicforever oh sorry, can you tell me more about it. I don't
understand well with what you mean. The
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330191884
Anyway if it helps you debug what is going on to help me:
I am using:
- MacBook Pro 13" mid 2014 model (i.e.
formath opened a new pull request #7932: executor not expose its inner
variables to symbol
URL: https://github.com/apache/incubator-mxnet/pull/7932
This is an automated message from the Apache Git Service.
To respond to
adrianloy commented on issue #6474: Fine-tune the mxnet ssd get
mismatchfrom.shape() error
URL:
https://github.com/apache/incubator-mxnet/issues/6474#issuecomment-330187205
When I had to do with this project, I also had to adjust it in demo.py and
when preparing the dataset. But I do
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330187016
because when trying something else
```
x = nd.ones(shape=(3,3))
x_gpu = x.copyto(gpu(0))
print(x_gpu)
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330186609
But the main issue is that it seems that your GPU command examples might not
be correct, given that the errors are
SumNeuron commented on issue #7900: Request: finish python gpu enabled guide
for install
URL:
https://github.com/apache/incubator-mxnet/issues/7900#issuecomment-330186373
also, I want to reiterate that I would appreciate it if you could implement
a gpu connection test
either of
formath closed pull request #7930: executor not expose its inner variables to
symbol
URL: https://github.com/apache/incubator-mxnet/pull/7930
This is an automated message from the Apache Git Service.
To respond to the
SumNeuron commented on issue #7917: SequenceLast of LSTM throws error
URL:
https://github.com/apache/incubator-mxnet/issues/7917#issuecomment-330184409
Perhaps you are seeing something that I am not.
This shows how to define blocks utilizing gluon which is the newer version
of symbol.
lijuan123 commented on issue #6474: Fine-tune the mxnet ssd get
mismatchfrom.shape() error
URL:
https://github.com/apache/incubator-mxnet/issues/6474#issuecomment-330156954
@adrianloy Hi , I have set the num_class and class_names in train.py, but i
still have the erro. So i want to know
ykim362 opened a new pull request #7931: MKL-DNN integration: request for
reviews
URL: https://github.com/apache/incubator-mxnet/pull/7931
**This PR is a beta version for the code reviews. There are several known
issues which are being debugged.**
# MKL-DNN
A new open-source
formath opened a new pull request #7930: executor not expose its inner
variables to symbol
URL: https://github.com/apache/incubator-mxnet/pull/7930
This is an automated message from the Apache Git Service.
To respond to
cjolivier01 commented on issue #7926: Increase the tolerance
URL: https://github.com/apache/incubator-mxnet/pull/7926#issuecomment-330143129
Was it failing before last week? Because if it wans t, it's not the fault
of the test. Not falling without 0.01 error is a problem.
On Sun,
feiyuvl commented on issue #7871: Why library size so large
URL:
https://github.com/apache/incubator-mxnet/issues/7871#issuecomment-330135336
No, just feel it is too large. A single obj file generated by some op is
about 10M which is unreasonable. Maybe the using of template in mshadow
feiyuvl commented on issue #7871: Why library size so large
URL:
https://github.com/apache/incubator-mxnet/issues/7871#issuecomment-330135336
No, just feel it is too large. A single obj file generated by some op is
about 10M which is unreasonable. Maybe the using of template in mshadow
feiyuvl commented on issue #7871: Why library size so large
URL:
https://github.com/apache/incubator-mxnet/issues/7871#issuecomment-330135336
No, just feel it is too large. A single obj file generated by some op is
above 10M which is unreasonable. Maybe the using of template in mshadow
96 matches
Mail list logo