wkcn commented on issue #16747: Fused Op causes MXNetError
URL:
https://github.com/apache/incubator-mxnet/issues/16747#issuecomment-550971205
I agree to turn the fused_op off until fused_op is stable, and use the
environmental variable to switch it.
The reason is that users couldn't use
knjwhn opened a new issue #16749: Ask for advice about using my int8gemm
URL: https://github.com/apache/incubator-mxnet/issues/16749
Hello everyone.
I implemented a gemm multiply function for u8s8s32 and s8s8s32 data types ,
and the function interface are same as openblas , I want to us
TaoLv commented on issue #16747: Fused Op causes MXNetError
URL:
https://github.com/apache/incubator-mxnet/issues/16747#issuecomment-550978412
+1
This is an automated message from the Apache Git Service.
To respond to the mes
knjwhn commented on issue #16749: Ask for advice about using my int8gemm
URL:
https://github.com/apache/incubator-mxnet/issues/16749#issuecomment-550979571
https://github.com/apache/incubator-mxnet/blob/c38b52784325380f79cafb9f0407ad327554fe6b/src/operator/quantization/quantized_fully_con
TaoLv commented on issue #16749: Ask for advice about using my int8gemm
URL:
https://github.com/apache/incubator-mxnet/issues/16749#issuecomment-550984158
Are you running experiments or targeting for upstreaming in the future? If
just experiments, I think you can directly replace the `cbl
knjwhn commented on issue #16749: Ask for advice about using my int8gemm
URL:
https://github.com/apache/incubator-mxnet/issues/16749#issuecomment-550989453
thanks ! I build MXNet with USE_BLAS=openblas , because of my target
machine is arm64 so I can't using mkl or mkldnn, and by avoidi
wkcn edited a comment on issue #16747: Fused Op causes MXNetError
URL:
https://github.com/apache/incubator-mxnet/issues/16747#issuecomment-550971205
I agree to turn the fused_op off by default until fused_op is stable.
The reason is that users couldn't use the 1.6.0 release if it is not
TaoLv commented on issue #16749: Ask for advice about using my int8gemm
URL:
https://github.com/apache/incubator-mxnet/issues/16749#issuecomment-550998461
I see. Given you already have INT8 GEMM, so I would suggest you start from
the normal FP32 FullyConnected operator and replace its imp
marcoabreu commented on issue #16412: Cleanup output of docker cache generation
URL: https://github.com/apache/incubator-mxnet/pull/16412#issuecomment-551013603
Thanks, I'd prefer to not surpress the output and not move forward with the
change
--
wuxun-zhang commented on issue #16184: Add large tensor nightly tests for
MKL-DNN operators
URL: https://github.com/apache/incubator-mxnet/pull/16184#issuecomment-551030416
Output log of all mkl-dnn operators (with `export MKLDNN_VERBOSE=1` ):
```
test_large_array_mkldnn.test_Full
knjwhn commented on issue #16749: Ask for advice about using my int8gemm
URL:
https://github.com/apache/incubator-mxnet/issues/16749#issuecomment-551031875
Thanks a lot !I will try that.and is there have some study resource of how
to use quantization/dequantization code? What's more I st
Alicia1529 opened a new pull request #16750: [Numpy] add custom op sort
URL: https://github.com/apache/incubator-mxnet/pull/16750
## Description ##
add custom NumPy operator sort
This is an automated message from the Ap
pengzhao-intel commented on issue #16731: [WIP] Static link MKL-DNN library
URL: https://github.com/apache/incubator-mxnet/pull/16731#issuecomment-551052887
@Taolv could you give a simple summary of the reason for this change and any
possible impacts?
--
This is an automated email from the ASF dual-hosted git repository.
aaronmarkham pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git
The following commit(s) were added to refs/heads/asf-site by this push:
new 1f2ed68 Bump the publis
linven0721 opened a new issue #16751: Run Java example, GPU, Memory leak
URL: https://github.com/apache/incubator-mxnet/issues/16751
## Description
When I tried to run the Java predictor
example([https://github.com/apache/incubator-mxnet/blob/master/scala-package/examples/src/main/java/o
TaoLv commented on issue #16731: [WIP] Static link MKL-DNN library
URL: https://github.com/apache/incubator-mxnet/pull/16731#issuecomment-551106435
@pengzhao-intel , sure, PR description is added.
This is an automated message
TaoLv commented on issue #16749: Ask for advice about using my int8gemm
URL:
https://github.com/apache/incubator-mxnet/issues/16749#issuecomment-551112964
MXNet INT8 Conv doesn't call GEMM directly. It leverages the Conv APIs from
MKL-DNN/cuDNN.
--
TaoLv commented on issue #16731: [WIP] Static link MKL-DNN library
URL: https://github.com/apache/incubator-mxnet/pull/16731#issuecomment-551126579
@szha @lanking520 @perdasilva @marcoabreu Could you please review,
especially the parts about CD and packaging?
--
JONGGON opened a new issue #16752: Gluon hybridize is not perfect.
URL: https://github.com/apache/incubator-mxnet/issues/16752
## Description
(A clear and concise description of what the bug is.)
I am implementing yolov3.
If hybridize is not active, no problem will occur.
Whe
reminisce commented on issue #16716: [Numpy] Fix collect_params().zero_grad()
in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551173088
I'm personally not fond of using the operator `reset_array` for some
specific purpose. First of all, it do
reminisce edited a comment on issue #16716: [Numpy] Fix
collect_params().zero_grad() in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551173088
I'm personally not fond of using the operator `reset_array` for some
specific purpose. First of all
alopez1327 opened a new issue #16753: fail to build using docker
URL: https://github.com/apache/incubator-mxnet/issues/16753
## Description
Tried building library using docker through the command: _python3
ci/build.py -p armv7_ but compilation failed because it seems docker is pulling
s
reminisce commented on a change in pull request #16750: [Numpy] add custom op
sort
URL: https://github.com/apache/incubator-mxnet/pull/16750#discussion_r343782233
##
File path: python/mxnet/ndarray/numpy/_op.py
##
@@ -344,6 +344,133 @@ def take(a, indices, axis=None, mode=
reminisce commented on a change in pull request #16750: [Numpy] add custom op
sort
URL: https://github.com/apache/incubator-mxnet/pull/16750#discussion_r343784245
##
File path: python/mxnet/numpy_op_fallback.py
##
@@ -49,6 +49,48 @@ def _register_helper(prop_cls):
ret
anirudh2290 commented on a change in pull request #16748: Fix SliceChannel Type
inference
URL: https://github.com/apache/incubator-mxnet/pull/16748#discussion_r343786519
##
File path: src/operator/slice_channel-inl.h
##
@@ -176,16 +177,22 @@ class SliceChannelProp : public
anirudh2290 commented on a change in pull request #16748: Fix SliceChannel Type
inference
URL: https://github.com/apache/incubator-mxnet/pull/16748#discussion_r343786780
##
File path: tests/python/gpu/test_contrib_amp.py
##
@@ -475,6 +475,15 @@ def test_fp16_casting():
Wallart opened a new issue #16754: Is mirroring working with MXNet 1.5.1 Gluon ?
URL: https://github.com/apache/incubator-mxnet/issues/16754
Hello everyone,
I am currently implementing a Sparse/LogSparse transformer.
I would like some layers results to be re-computed at backward pass t
sxjscience commented on issue #16576: [Numpy][Bug] einsum bug
URL:
https://github.com/apache/incubator-mxnet/issues/16576#issuecomment-551203792
I've rerun the script and can confirm that @hzfan's commit has solved the
problem. I'll later try to use einsum in the implementation of neural a
sxjscience closed issue #16576: [Numpy][Bug] einsum bug
URL: https://github.com/apache/incubator-mxnet/issues/16576
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
sxjscience commented on issue #16752: Gluon hybridize is not perfect.
URL:
https://github.com/apache/incubator-mxnet/issues/16752#issuecomment-551205420
@JONGGON There are a few know inconsistent behaviors the HybridBlock:
https://github.com/apache/incubator-mxnet/issues/16279,
https://gi
sxjscience edited a comment on issue #16752: Gluon hybridize is not perfect.
URL:
https://github.com/apache/incubator-mxnet/issues/16752#issuecomment-551205420
@JONGGON There are a few know inconsistent behaviors the HybridBlock:
https://github.com/apache/incubator-mxnet/issues/16279,
htt
sxjscience commented on issue #12268: Inconsistent type conversion from
numpy.ndarray to mx.ndarray
URL:
https://github.com/apache/incubator-mxnet/issues/12268#issuecomment-551207085
This issue is solved by the new Numpy NDArray in MXNet:
```python
import mxnet as mx
import numpy
sxjscience closed issue #3660: [OP] Convention for the `axis` argument
URL: https://github.com/apache/incubator-mxnet/issues/3660
This is an automated message from the Apache Git Service.
To respond to the message, please log
sxjscience closed issue #12268: Inconsistent type conversion from numpy.ndarray
to mx.ndarray
URL: https://github.com/apache/incubator-mxnet/issues/12268
This is an automated message from the Apache Git Service.
To respond t
sxjscience commented on issue #3822: [OP] index_fill and index_add operators
URL:
https://github.com/apache/incubator-mxnet/issues/3822#issuecomment-551208268
The new numpy ndarray supports this feature.
This is an automated
sxjscience closed issue #3822: [OP] index_fill and index_add operators
URL: https://github.com/apache/incubator-mxnet/issues/3822
This is an automated message from the Apache Git Service.
To respond to the message, please log
This is an automated email from the ASF dual-hosted git repository.
aaronmarkham pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git
The following commit(s) were added to refs/heads/asf-site by this push:
new 02d2b16 Bump the publis
reminisce commented on a change in pull request #16750: [Numpy] add custom op
sort
URL: https://github.com/apache/incubator-mxnet/pull/16750#discussion_r343814159
##
File path: python/mxnet/numpy/multiarray.py
##
@@ -2276,6 +2276,127 @@ def take(a, indices, axis=None, mode
alopez1327 commented on issue #16753: fail to build using docker
URL:
https://github.com/apache/incubator-mxnet/issues/16753#issuecomment-551216461
Tried to fix the error by adding:
_apt update && apt-key update && apt install -y --no-install-recommends_
to deb.ubuntu.ccache.sh be
reminisce commented on a change in pull request #16660: [WIP] [Numpy] TVM
implementation for binary ops
URL: https://github.com/apache/incubator-mxnet/pull/16660#discussion_r343827251
##
File path: contrib/tvmop/core/umath.py
##
@@ -120,3 +121,327 @@ def _compute_binary_sc
ptrendx commented on issue #16716: [Numpy] Fix collect_params().zero_grad() in
gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551230288
@reminisce No. That is not how it works and please do not undo performance
optimizations because of such fal
sxjscience commented on issue #16716: [Numpy] Fix collect_params().zero_grad()
in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551231994
@ptrendx I think you have misunderstood my comment. What I means is that we
could try to fuse these zeroi
ptrendx commented on issue #16747: Fused Op causes MXNetError
URL:
https://github.com/apache/incubator-mxnet/issues/16747#issuecomment-551232209
Isn't right now the period of finding those integration bugs and fixing them
for 1.6 release? I will definitely look into this issue and fix it,
ptrendx commented on issue #16716: [Numpy] Fix collect_params().zero_grad() in
gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551233612
@sxjscience How would you like to fuse them without writing a new operator?
sxjscience commented on issue #16747: Fused Op causes MXNetError
URL:
https://github.com/apache/incubator-mxnet/issues/16747#issuecomment-551235898
@ptrendx I think we are already in a code-freeze status and the simplest fix
is to turn it off by default. We could easily turn it on in 1.6.1
sxjscience commented on issue #16716: [Numpy] Fix collect_params().zero_grad()
in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551237107
@ptrendx For example, we can later try to hybridize part of the imperative
codes by the nvrtc approach th
sxjscience commented on issue #16716: [Numpy] Fix collect_params().zero_grad()
in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551237737
@ptrendx The problem of the `reset_arrays` approach is that the users may
later require `fill_arrays`, wh
access2rohit commented on issue #16737: [MKLDNN] use dim_t instead of int in
slice/transpose operators
URL: https://github.com/apache/incubator-mxnet/pull/16737#issuecomment-551240748
@wuxun-zhang thank you for your contribution and providing a quick fix.
-
reminisce commented on issue #16716: [Numpy] Fix collect_params().zero_grad()
in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551241818
@ptrendx The the performance overhead in your benchmark really comes from
the FFI and pushing ops to the a
sxjscience commented on issue #16716: [Numpy] Fix collect_params().zero_grad()
in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551243605
> scare me, frankly. Doing numpy-like fully imperative execution has no
chance of being actually fast and
reminisce commented on issue #16716: [Numpy] Fix collect_params().zero_grad()
in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551246891
Another big factor that may contribute to the slowdown of assigning zeros is
through `a[:] = 0` which has
ChaiBapchya opened a new pull request #16755: Enabling large tensor support for
binary broadcast operators
URL: https://github.com/apache/incubator-mxnet/pull/16755
## Description ##
In continuation of #16714
Add tests
- arctan2
- hypot
## Checklist ##
### Essentials
ptrendx commented on issue #16716: [Numpy] Fix collect_params().zero_grad() in
gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551264243
Ok, let me address those comments 1 point at a time :-).
- usage of TVM/nvrtc - I am generally in favor o
access2rohit commented on issue #16732: MKLDNN-1.0 doesn't support slice
operator for Large Tensor
URL:
https://github.com/apache/incubator-mxnet/issues/16732#issuecomment-551276027
PR for the fix: https://github.com/apache/incubator-mxnet/pull/16737
--
access2rohit edited a comment on issue #16732: MKLDNN-1.0 doesn't support slice
operator for Large Tensor
URL:
https://github.com/apache/incubator-mxnet/issues/16732#issuecomment-551276027
PR: https://github.com/apache/incubator-mxnet/pull/16737
fixes the issue
access2rohit commented on issue #16737: [MKLDNN] use dim_t instead of int in
slice/transpose operators
URL: https://github.com/apache/incubator-mxnet/pull/16737#issuecomment-551276539
@ChaiBapchya can you also verify ?
This i
CanyonWind commented on issue #16424: [Channel Shuffle / Hard Swish / Hard
Sigmoid] running in MKL CPU backend failed
URL:
https://github.com/apache/incubator-mxnet/issues/16424#issuecomment-551277963
Hi @ZhennanQin, thanks a lot to your effort!
I tried to verify the quantized model's
ptrendx commented on issue #16747: Fused Op causes MXNetError
URL:
https://github.com/apache/incubator-mxnet/issues/16747#issuecomment-551284480
Ok, I sent a clarification email to dev@ as you are not actually the first
person to reach out to me with this misunderstanding of code freeze. C
zachgk commented on issue #16560: It is easy to crash MXNet when tensor goes
larger
URL:
https://github.com/apache/incubator-mxnet/issues/16560#issuecomment-551285938
Is this resolved now that #16570 is merged?
This is an au
sxjscience commented on issue #16716: [Numpy] Fix collect_params().zero_grad()
in gluon numpy interface
URL: https://github.com/apache/incubator-mxnet/pull/16716#issuecomment-551291281
@ptrendx
Let me clarify a little bit:
1) In the nd interface, using `reset_arrays` as a short-t
CanyonWind edited a comment on issue #16424: [Channel Shuffle / Hard Swish /
Hard Sigmoid] running in MKL CPU backend failed
URL:
https://github.com/apache/incubator-mxnet/issues/16424#issuecomment-551277963
Hi @ZhennanQin, thanks a lot for your effort!
I tried to verify the quantized
leezu commented on issue #16747: Fused Op causes MXNetError
URL:
https://github.com/apache/incubator-mxnet/issues/16747#issuecomment-551312428
I agree with @ptrendx, we should try to fix the bugs and ship the features
if time allows.
---
sxjscience commented on issue #16747: Fused Op causes MXNetError
URL:
https://github.com/apache/incubator-mxnet/issues/16747#issuecomment-551316009
I received the clarification email about the meaning of code freeze and I
agree with @ptrendx that we should try to fix it these days and cons
anirudh2290 opened a new pull request #16756: [WIP] Multithreaded inference
backend support
URL: https://github.com/apache/incubator-mxnet/pull/16756
## Description ##
Trying to run CI against 1.6
## Checklist ##
### Essentials ###
Please feel free to remove inapplicable ite
access2rohit commented on issue #16755: Enabling large tensor support for
binary broadcast operators
URL: https://github.com/apache/incubator-mxnet/pull/16755#issuecomment-551323548
LGTM ! Also paste the output for tests run.
Can we also do large vector tests too for arctan2 ? I assume h
ptrendx commented on issue #16748: Fix SliceChannel Type inference
URL: https://github.com/apache/incubator-mxnet/pull/16748#issuecomment-551323521
I don't know enough to comment on the changes to the example etc., but the
main fix from this PR (InferType in SliceChannel) looks good. @aniru
anirudh2290 edited a comment on issue #16748: Fix SliceChannel Type inference
URL: https://github.com/apache/incubator-mxnet/pull/16748#issuecomment-551323980
@ptrendx how did you come up with that sentence ?
This is an automa
anirudh2290 commented on issue #16748: Fix SliceChannel Type inference
URL: https://github.com/apache/incubator-mxnet/pull/16748#issuecomment-551323980
@ptrendx how did you come up with that regex ?
This is an automated messag
anirudh2290 edited a comment on issue #16748: Fix SliceChannel Type inference
URL: https://github.com/apache/incubator-mxnet/pull/16748#issuecomment-551323980
@ptrendx how did you come up with that sentence ?
EDIT: Never mind, its the same error message.
-
anirudh2290 commented on issue #16748: Fix SliceChannel Type inference
URL: https://github.com/apache/incubator-mxnet/pull/16748#issuecomment-551326529
@ptrendx good point about the other places where error is raised. The scope
of the PR increased :) . Can we start with this PR first and I
ChaiBapchya commented on issue #16755: Enabling large tensor support for binary
broadcast operators
URL: https://github.com/apache/incubator-mxnet/pull/16755#issuecomment-551328930
Build flag
```
python -c "from mxnet.runtime import feature_list; print(feature_list())"
[✖ CUDA, ✖ C
This is an automated email from the ASF dual-hosted git repository.
aaronmarkham pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet-site.git
The following commit(s) were added to refs/heads/asf-site by this push:
new 961e79d Bump the publis
DickJC123 commented on issue #15167: Pointwise fusion for GPU
URL: https://github.com/apache/incubator-mxnet/pull/15167#issuecomment-551332574
After some investigation, I have an explanation and planned fix for the perf
regression. To repeat what @ptrendx mentions, the real-time compilation
DickJC123 edited a comment on issue #15167: Pointwise fusion for GPU
URL: https://github.com/apache/incubator-mxnet/pull/15167#issuecomment-551332574
After some investigation, I have an explanation and planned fix for the perf
regression. To repeat what @ptrendx mentions, the real-time comp
larroy commented on issue #16412: Cleanup output of docker cache generation
URL: https://github.com/apache/incubator-mxnet/pull/16412#issuecomment-551334777
Thanks for your review. So going forward will you maintain this
infrastructure then?
anirudh2290 opened a new issue #16757: Exceptions thrown in InferType before
doing reverse inference
URL: https://github.com/apache/incubator-mxnet/issues/16757
## Description
Some operators are missing logic for reverse inference: inferring the type
of inputs from outputs. Instead, exc
wuxun-zhang commented on issue #16732: MKLDNN-1.0 doesn't support slice
operator for Large Tensor
URL:
https://github.com/apache/incubator-mxnet/issues/16732#issuecomment-551337898
Glad it works.
This is an automated messag
pengzhao-intel commented on issue #16731: [WIP] Static link MKL-DNN library
URL: https://github.com/apache/incubator-mxnet/pull/16731#issuecomment-551341109
> @pengzhao-intel , sure, PR description is added.
Very great explanation. Ping me when your PR is ready to mer
pengzhao-intel merged pull request #16737: [MKLDNN] use dim_t instead of int in
slice/transpose operators
URL: https://github.com/apache/incubator-mxnet/pull/16737
This is an automated message from the Apache Git Service.
To
pengzhao-intel commented on issue #16737: [MKLDNN] use dim_t instead of int in
slice/transpose operators
URL: https://github.com/apache/incubator-mxnet/pull/16737#issuecomment-551343510
I am merging this PR first and feel free to ping us if other issues are
found.
---
This is an automated email from the ASF dual-hosted git repository.
patriczhao pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from c38b527 [MKLDNN] Fix int8 convolution/fc bias overflow (#16734)
add 5dfa121 [MKLDNN] use dim_t in
This is an automated email from the ASF dual-hosted git repository.
patriczhao pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from c38b527 [MKLDNN] Fix int8 convolution/fc bias overflow (#16734)
add 5dfa121 [MKLDNN] use dim_t in
pengzhao-intel commented on issue #16749: Ask for advice about using my
int8gemm
URL:
https://github.com/apache/incubator-mxnet/issues/16749#issuecomment-551345467
@knjwhn Does current int8 FC OP work for you?
Could you provide some backgrounds of your workload and target with int
pengzhao-intel commented on issue #16749: Ask for advice about using my
int8gemm
URL:
https://github.com/apache/incubator-mxnet/issues/16749#issuecomment-551346152
More info in our blog:
https://medium.com/apache-mxnet/model-quantization-for-production-level-neural-network-inference-f54
hzfan commented on a change in pull request #16660: [WIP] [Numpy] TVM
implementation for binary ops
URL: https://github.com/apache/incubator-mxnet/pull/16660#discussion_r343961440
##
File path: contrib/tvmop/core/umath.py
##
@@ -120,3 +121,327 @@ def _compute_binary_scalar
ptrendx commented on issue #16748: Fix SliceChannel Type inference
URL: https://github.com/apache/incubator-mxnet/pull/16748#issuecomment-551352067
Sure, having this PR fix just SliceChannel is totally fine.
This is an automat
leezu closed issue #16753: fail to build using docker
URL: https://github.com/apache/incubator-mxnet/issues/16753
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub an
leezu commented on issue #16753: fail to build using docker
URL:
https://github.com/apache/incubator-mxnet/issues/16753#issuecomment-551354006
I believe `ci/build.py` is not suppported outside of the CI environment. So
I'm closing the issue.
Do you have any issues with cross-compile?
-
leezu edited a comment on issue #16753: fail to build using docker
URL:
https://github.com/apache/incubator-mxnet/issues/16753#issuecomment-551354006
I believe `ci/build.py` is not suppported outside of the CI environment. So
I'm closing the issue. Feel free to reopen if you believe it is
vexilligera commented on issue #16730: [NumPy] NumPy support for linalg.inv
URL: https://github.com/apache/incubator-mxnet/pull/16730#issuecomment-551354516
@reminisce @haojin2
This is an automated message from the Apache Git
leezu opened a new issue #16758: [Estimator] Emits warnings when mixing
user-defined and default event handlers
URL: https://github.com/apache/incubator-mxnet/issues/16758
## Description
Estimator emits a arguably useless warning, when users specify
`event_handlers`.
### Error Me
anirudh2290 merged pull request #16748: Fix SliceChannel Type inference
URL: https://github.com/apache/incubator-mxnet/pull/16748
This is an automated message from the Apache Git Service.
To respond to the message, please log
This is an automated email from the ASF dual-hosted git repository.
anirudh2290 pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 5dfa121 [MKLDNN] use dim_t instead of int in slice/transpose
operators (#16737)
add a37dcd4 Fix
ChaiBapchya commented on issue #16755: Enabling large tensor support for binary
broadcast operators
URL: https://github.com/apache/incubator-mxnet/pull/16755#issuecomment-551359931
> LGTM ! Also paste the output for tests run.
> Can we also do large vector tests too for arctan2 ? I a
ChaiBapchya commented on issue #16755: Enabling large tensor support for binary
broadcast operators
URL: https://github.com/apache/incubator-mxnet/pull/16755#issuecomment-551360769
Vector test
```
nosetests tests/nightly/test_large_vector.py:test_binary_broadcast
.
-
TaoLv commented on issue #11417: libomp.so dependency (need REAL fix)
URL:
https://github.com/apache/incubator-mxnet/issues/11417#issuecomment-551361750
> btw there's a problem with the mkldnn build. It pulls in libgomp always:
@cjolivier01 It seems that even MXNet is not built with
zhly0 opened a new issue #16759: LeakyReLU
URL: https://github.com/apache/incubator-mxnet/issues/16759
## Description
(in the doc:
http://mxnet.incubator.apache.org/api/python/docs/api/symbol/op/index.html#mxnet.symbol.op.LeakyReLU.it
is said that leakyrelu is activate on input elem
TaoLv edited a comment on issue #11417: libomp.so dependency (need REAL fix)
URL:
https://github.com/apache/incubator-mxnet/issues/11417#issuecomment-551361750
> btw there's a problem with the mkldnn build. It pulls in libgomp always:
@cjolivier01 It seems that even MXNet is not buil
TaoLv edited a comment on issue #11417: libomp.so dependency (need REAL fix)
URL:
https://github.com/apache/incubator-mxnet/issues/11417#issuecomment-551361750
> btw there's a problem with the mkldnn build. It pulls in libgomp always:
@cjolivier01 It seems that even MXNet is not buil
leezu closed issue #16758: [Estimator] Emits warnings when mixing user-defined
and default event handlers
URL: https://github.com/apache/incubator-mxnet/issues/16758
This is an automated message from the Apache Git Service.
1 - 100 of 112 matches
Mail list logo