Zha0q1 commented on a change in pull request #15490: Utility to help developers
debug operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#discussion_r303082695
##
File path: src/common/tensor_inspector.h
##
@@ -0,0 +1,733 @@
+/*
+ *
Zha0q1 commented on a change in pull request #15490: Utility to help developers
debug operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#discussion_r303082645
##
File path: src/common/tensor_inspector.h
##
@@ -0,0 +1,733 @@
+/*
+ *
Zha0q1 commented on a change in pull request #15490: Utility to help developers
debug operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#discussion_r303081013
##
File path: src/common/tensor_inspector.h
##
@@ -0,0 +1,733 @@
+/*
+ *
Zha0q1 commented on a change in pull request #15490: Utility to help developers
debug operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#discussion_r303081013
##
File path: src/common/tensor_inspector.h
##
@@ -0,0 +1,733 @@
+/*
+ *
Zha0q1 commented on a change in pull request #15490: Utility to help developers
debug operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#discussion_r303073483
##
File path: src/common/tensor_inspector.h
##
@@ -0,0 +1,733 @@
+/*
+ *
haojin2 merged pull request #15302: [Numpy] Numpy hstack
URL: https://github.com/apache/incubator-mxnet/pull/15302
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
This is an automated email from the ASF dual-hosted git repository.
haoj pushed a commit to branch numpy
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/numpy by this push:
new a26b201 [Numpy] Numpy hstack (#15302)
This is an automated email from the ASF dual-hosted git repository.
kellen pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 554b196 Rebase #13757 to master (#15189)
add 6a564be cuda/cuDNN lib version checking. Force cuDNN
KellenSunderland merged pull request #15449: cuda/cuDNN lib version checking.
Force cuDNN v7 usage.
URL: https://github.com/apache/incubator-mxnet/pull/15449
This is an automated message from the Apache Git Service.
To
KellenSunderland commented on a change in pull request #15449: cuda/cuDNN lib
version checking. Force cuDNN v7 usage.
URL: https://github.com/apache/incubator-mxnet/pull/15449#discussion_r303065500
##
File path: src/common/cuda_utils.cc
##
@@ -0,0 +1,108 @@
+/*
+ *
TaoLv commented on issue #15521: Avoid memory copy for dropout inference
URL: https://github.com/apache/incubator-mxnet/pull/15521#issuecomment-510944094
@eric-haibin-lin please help to review.
This is an automated message
ChaiBapchya commented on issue #15506: Improving error message
URL:
https://github.com/apache/incubator-mxnet/issues/15506#issuecomment-510944105
Alright. If it is tough to safely tell the location, can't it be renamed to
```
fatal error : No space left on device
compilation
TaoLv opened a new pull request #15521: Avoid memory copy for dropout inference
URL: https://github.com/apache/incubator-mxnet/pull/15521
## Description ##
Fix the first item in #14198.
## Checklist ##
### Essentials ###
Please feel free to remove inapplicable items for
frankfliu commented on issue #15520: most time cost in NDArray GetData() func
URL:
https://github.com/apache/incubator-mxnet/issues/15520#issuecomment-510943143
@mxnet-label-bot add [question, performance]
This is an
braindotai commented on issue #12185: from_logits definition seems different
from what is expected?
URL:
https://github.com/apache/incubator-mxnet/issues/12185#issuecomment-510939653
[Check this
out](https://www.tensorflow.org/api_docs/python/tf/nn/softmax_cross_entropy_with_logits)
smissan closed issue #15514: dmlc.lib is not generated
URL: https://github.com/apache/incubator-mxnet/issues/15514
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
mikemwx commented on issue #15302: [Numpy] Numpy hstack
URL: https://github.com/apache/incubator-mxnet/pull/15302#issuecomment-510876247
> LGTM, remember to add examples to the doc and it will be good for merge
Docs are added
gxkevin opened a new issue #15520: most time cost in NDArray GetData() func
URL: https://github.com/apache/incubator-mxnet/issues/15520
My predict code is like this, I have 32 worker thread, each has 1 openmp
thread, the mxnet engine is NaiveEngine , using the openblas, disable the
KhurramPirov commented on issue #15496: Loaded pretrained model but train
accuracy starts from zero,
URL:
https://github.com/apache/incubator-mxnet/issues/15496#issuecomment-510873847
It doesn't work inside batch_callback and model.fit.
пт, 12 июл. 2019 г. в 14:08, Rishik Mourya :
KhurramPirov commented on issue #15504: Saved model and then looked at the
output of the same batch after loading, got different results!
URL:
https://github.com/apache/incubator-mxnet/issues/15504#issuecomment-510873701
It doesn't work inside batch_callback.
пт, 12 июл. 2019 г. в
pengzhao-intel edited a comment on issue #13598: More fine-grained operator
implementation dispatch & memory planning flow
URL:
https://github.com/apache/incubator-mxnet/issues/13598#issuecomment-472725241
@danithaca Tao has answered the question in the forum :) thanks to raising
the
wkcn edited a comment on issue #15499: Improve diagnose.py, adding build
features info and binary library path.
URL: https://github.com/apache/incubator-mxnet/pull/15499#issuecomment-510861117
Thank you for the improvement! LGTM.
wkcn commented on issue #15499: Improve diagnose.py, adding build features info
and binary library path.
URL: https://github.com/apache/incubator-mxnet/pull/15499#issuecomment-510861117
Thank you for the improvement!
I have a suggestion : )
Could we only print the relative path?
marcoabreu commented on issue #15506: Improving error message
URL:
https://github.com/apache/incubator-mxnet/issues/15506#issuecomment-510860930
I think that this error is a quite standard one. Considering that the
compiler might write to a lot of locations due to optimizations, I don't
apeforest commented on issue #15490: Utility to help developers debug
operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#issuecomment-510847250
@eric-haibin-lin This utility looks really useful for debugging mxnet
backend. Please take a look if
braindotai commented on issue #15496: Loaded pretrained model but train
accuracy starts from zero,
URL:
https://github.com/apache/incubator-mxnet/issues/15496#issuecomment-510847042
Below code should work
```python
model.save_checkpoint('model', epoch = 0)
sym, arg_params,
apeforest commented on a change in pull request #15490: Utility to help
developers debug operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#discussion_r302935550
##
File path: src/common/tensor_inspector.h
##
@@ -0,0 +1,733 @@
+/*
+ *
apeforest commented on a change in pull request #15490: Utility to help
developers debug operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#discussion_r302934098
##
File path: src/common/tensor_inspector.h
##
@@ -0,0 +1,733 @@
+/*
+ *
apeforest commented on a change in pull request #15490: Utility to help
developers debug operators: Tensor Inspector
URL: https://github.com/apache/incubator-mxnet/pull/15490#discussion_r302933381
##
File path: src/common/tensor_inspector.h
##
@@ -0,0 +1,733 @@
+/*
+ *
haojin2 commented on a change in pull request #15169: Softmax with length
URL: https://github.com/apache/incubator-mxnet/pull/15169#discussion_r302932168
##
File path: src/operator/nn/softmax-inl.h
##
@@ -464,35 +755,56 @@ void SoftmaxGradCompute(const nnvm::NodeAttrs&
haojin2 commented on a change in pull request #15169: Softmax with length
URL: https://github.com/apache/incubator-mxnet/pull/15169#discussion_r302931553
##
File path: tests/python/unittest/test_operator.py
##
@@ -7838,6 +7871,7 @@ def get_output_names_callback(name, arr):
haojin2 commented on a change in pull request #15169: Softmax with length
URL: https://github.com/apache/incubator-mxnet/pull/15169#discussion_r302931319
##
File path: src/operator/nn/softmax-inl.h
##
@@ -464,35 +755,56 @@ void SoftmaxGradCompute(const nnvm::NodeAttrs&
haojin2 commented on a change in pull request #15169: Softmax with length
URL: https://github.com/apache/incubator-mxnet/pull/15169#discussion_r302931406
##
File path: src/operator/nn/softmax-inl.h
##
@@ -113,6 +113,60 @@ inline void Softmax(Stream *s, DType *in, OType
haojin2 commented on issue #15169: Softmax with length
URL: https://github.com/apache/incubator-mxnet/pull/15169#issuecomment-510842258
@apeforest Trust me or not, I've tried what you're suggesting to do on my
end way before this PR was raised but found it not easily done without
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302928682
##
File path: tests/python/gpu/test_operator_gpu.py
##
@@ -2300,6 +2300,81 @@ def test_math():
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302928767
##
File path: tests/python/gpu/test_operator_gpu.py
##
@@ -2300,6 +2300,81 @@ def test_math():
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302928820
##
File path: src/operator/numpy/np_elemwise_broadcast_op.cu
##
@@ -42,6 +42,12 @@
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302928599
##
File path: tests/python/gpu/test_operator_gpu.py
##
@@ -2300,6 +2300,81 @@ def test_math():
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302928018
##
File path: src/operator/numpy/np_elemwise_broadcast_op.cc
##
@@ -198,5 +198,74 @@
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302927937
##
File path: src/operator/numpy/np_elemwise_broadcast_op.cc
##
@@ -198,5 +198,74 @@
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302927764
##
File path: src/operator/numpy/np_elemwise_broadcast_op.cc
##
@@ -198,5 +198,74 @@
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302927529
##
File path: src/operator/numpy/np_elemwise_broadcast_op.cc
##
@@ -198,5 +198,74 @@
haojin2 commented on a change in pull request #15519: [Numpy] arctan2, a new
branch for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519#discussion_r302927677
##
File path: src/operator/numpy/np_elemwise_broadcast_op.cc
##
@@ -198,5 +198,74 @@
qqaatw edited a comment on issue #15484: Binding Model fails with simple_bind
error
URL:
https://github.com/apache/incubator-mxnet/issues/15484#issuecomment-510480630
@Maicus I have the same problem as you, could you tell more in detail about
the Target and the TargetCode Section?
haojin2 commented on a change in pull request #15468: Numpy compatible diagflat
operator
URL: https://github.com/apache/incubator-mxnet/pull/15468#discussion_r302922300
##
File path: src/operator/numpy/np_diagflat_op.cu
##
@@ -0,0 +1,38 @@
+/*
+ * Licensed to the Apache
haojin2 commented on a change in pull request #15468: Numpy compatible diagflat
operator
URL: https://github.com/apache/incubator-mxnet/pull/15468#discussion_r302921968
##
File path: tests/python/unittest/test_numpy_op.py
##
@@ -28,6 +28,59 @@
import random
haojin2 commented on a change in pull request #15468: Numpy compatible diagflat
operator
URL: https://github.com/apache/incubator-mxnet/pull/15468#discussion_r302921755
##
File path: src/operator/numpy/np_diagflat_op-inl.h
##
@@ -0,0 +1,201 @@
+/*
Review comment:
This is an automated email from the ASF dual-hosted git repository.
haoj pushed a change to branch numpy
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 19c7c6e fix memory override bug in multinomial (#15397)
add 3ea5f8c numpy eye op (#15282)
No new
haojin2 merged pull request #15282: Numpy compatible eye
URL: https://github.com/apache/incubator-mxnet/pull/15282
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
haojin2 commented on a change in pull request #15468: Numpy compatible diagflat
operator
URL: https://github.com/apache/incubator-mxnet/pull/15468#discussion_r302921428
##
File path: src/operator/numpy/np_diagflat_op-inl.h
##
@@ -0,0 +1,201 @@
+/*
Review comment:
KhurramPirov commented on issue #15504: Saved model and then looked at the
output of the same batch after loading, got different results!
URL:
https://github.com/apache/incubator-mxnet/issues/15504#issuecomment-510831907
I checked the path, I do it inside batch_callback and model.fit, can
braindotai commented on issue #15504: Saved model and then looked at the output
of the same batch after loading, got different results!
URL:
https://github.com/apache/incubator-mxnet/issues/15504#issuecomment-510829624
Make sure you are saving and loading the model from the same path.
tingying2020 opened a new pull request #15519: [Numpy] arctan2, a new branch
for pr.
URL: https://github.com/apache/incubator-mxnet/pull/15519
## Description ##
(Brief description on what this PR is about)
## Checklist ##
### Essentials ###
Please feel free to remove
tingying2020 closed pull request #15350: Add operator arctan2 with its test code
URL: https://github.com/apache/incubator-mxnet/pull/15350
This is an automated message from the Apache Git Service.
To respond to the message,
zoeygxy commented on a change in pull request #15346: numpy_split support param
type fix; numpy_vsplit added
URL: https://github.com/apache/incubator-mxnet/pull/15346#discussion_r302902846
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -1835,6 +1835,83 @@ def
mikemwx commented on a change in pull request #15456: [Numpy] numpy compatible
invert
URL: https://github.com/apache/incubator-mxnet/pull/15456#discussion_r302875177
##
File path: python/mxnet/ndarray/numpy/_op.py
##
@@ -1414,6 +1413,48 @@ def rint(x, out=None, **kwargs):
mikemwx commented on a change in pull request #15314: [Numpy] Numpy dstack
URL: https://github.com/apache/incubator-mxnet/pull/15314#discussion_r302874471
##
File path: src/operator/numpy/np_matrix_op.cc
##
@@ -55,6 +55,67 @@ bool NumpyTransposeShape(const nnvm::NodeAttrs&
mikemwx commented on a change in pull request #15302: [Numpy] Numpy hstack
URL: https://github.com/apache/incubator-mxnet/pull/15302#discussion_r302873300
##
File path: tests/python/unittest/test_numpy_op.py
##
@@ -860,6 +860,72 @@ def get_new_shape(shape, axis):
zixuanweeei commented on issue #15497: [MKLDNN] Independent gradients requests
check with respect to weights and bias of convolution
URL: https://github.com/apache/incubator-mxnet/pull/15497#issuecomment-510794204
@matteosal Thanks. That's great. I am working on a unit test for this
matteosal commented on issue #15497: [MKLDNN] Independent gradients requests
check with respect to weights and bias of convolution
URL: https://github.com/apache/incubator-mxnet/pull/15497#issuecomment-510787762
> @matteosal I tested your example with commit
roywei commented on issue #15452: fix nightly CI failure
URL: https://github.com/apache/incubator-mxnet/pull/15452#issuecomment-510776556
I have verified test_estimator passed. but test tutorials is still failing
due to other tutorials are failing. This change is needed nonetheless
```
ckt624 commented on a change in pull request #15349: Numpy Tensordot Operator
URL: https://github.com/apache/incubator-mxnet/pull/15349#discussion_r302852496
##
File path: tests/python/unittest/test_numpy_op.py
##
@@ -26,6 +26,151 @@
from mxnet.test_utils import
ckt624 commented on a change in pull request #15349: Numpy Tensordot Operator
URL: https://github.com/apache/incubator-mxnet/pull/15349#discussion_r302849369
##
File path: src/operator/numpy/np_tensordot_op-inl.h
##
@@ -0,0 +1,575 @@
+/*
+ * Licensed to the Apache
ckt624 commented on a change in pull request #15349: Numpy Tensordot Operator
URL: https://github.com/apache/incubator-mxnet/pull/15349#discussion_r302849525
##
File path: src/operator/numpy/np_tensordot_op-inl.h
##
@@ -0,0 +1,575 @@
+/*
+ * Licensed to the Apache
ckt624 commented on a change in pull request #15349: Numpy Tensordot Operator
URL: https://github.com/apache/incubator-mxnet/pull/15349#discussion_r302849247
##
File path: src/operator/numpy/np_tensordot_op-inl.h
##
@@ -0,0 +1,575 @@
+/*
+ * Licensed to the Apache
ckt624 commented on a change in pull request #15349: Numpy Tensordot Operator
URL: https://github.com/apache/incubator-mxnet/pull/15349#discussion_r302844970
##
File path: src/operator/numpy/np_tensordot_op.cc
##
@@ -0,0 +1,344 @@
+/*
+ * Licensed to the Apache Software
101 - 166 of 166 matches
Mail list logo