aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271878736
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/README.md
##
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271876944
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/BertQA.java
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271875597
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/BertDataParser.ja
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271876126
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/BertDataParser.ja
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271876017
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/BertDataParser.ja
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271879514
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/README.md
##
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271880386
##
File path:
scala-package/examples/src/test/java/org/apache/mxnetexamples/javaapi/infer/predictor/BertExampleT
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271877673
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/README.md
##
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271878149
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/README.md
##
aaronmarkham commented on a change in pull request #14592: Add BERT QA
Scala/Java example
URL: https://github.com/apache/incubator-mxnet/pull/14592#discussion_r271880109
##
File path:
scala-package/examples/src/main/java/org/apache/mxnetexamples/javaapi/infer/bert/README.md
##
lanking520 commented on issue #14610: [Feature Request] Disable lazy evaluation
URL:
https://github.com/apache/incubator-mxnet/issues/14610#issuecomment-479608403
@wkcn So it seemed the concept of Lazy evaluation is not totally applicable
to all ops based on the fact that op get executed e
This is an automated email from the ASF dual-hosted git repository.
haibin 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 53743d6 set _scale in Trainer using op
eric-haibin-lin merged pull request #14593: Set _scale in Trainer using
Optimizer rescale_grad
URL: https://github.com/apache/incubator-mxnet/pull/14593
This is an automated message from the Apache Git Service.
To respond to
This is an automated email from the ASF dual-hosted git repository.
aaronmarkham 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 28fe552 Updated documentation ab
aaronmarkham merged pull request #14493: Updated documentation about nightly
tests
URL: https://github.com/apache/incubator-mxnet/pull/14493
This is an automated message from the Apache Git Service.
To respond to the message
aaronmarkham commented on a change in pull request #14534: [DOC] Update ubuntu
install instructions from source
URL: https://github.com/apache/incubator-mxnet/pull/14534#discussion_r271872150
##
File path: docs/install/ubuntu_setup.md
##
@@ -140,9 +140,30 @@ On Ubuntu vers
aaronmarkham commented on a change in pull request #14534: [DOC] Update ubuntu
install instructions from source
URL: https://github.com/apache/incubator-mxnet/pull/14534#discussion_r271872562
##
File path: docs/install/ubuntu_setup.md
##
@@ -140,9 +140,30 @@ On Ubuntu vers
aaronmarkham commented on a change in pull request #14534: [DOC] Update ubuntu
install instructions from source
URL: https://github.com/apache/incubator-mxnet/pull/14534#discussion_r271873084
##
File path: docs/install/ubuntu_setup.md
##
@@ -140,9 +140,30 @@ On Ubuntu vers
aaronmarkham commented on a change in pull request #14534: [DOC] Update ubuntu
install instructions from source
URL: https://github.com/apache/incubator-mxnet/pull/14534#discussion_r271873910
##
File path: docs/install/ubuntu_setup.md
##
@@ -170,44 +191,72 @@ If building o
haohuanw commented on issue #14591: Reference engine from chunk via weak pointer
URL: https://github.com/apache/incubator-mxnet/pull/14591#issuecomment-479599564
LGTM
This is an automated message from the Apache Git Service.
T
szha commented on issue #14438: Fixing unintentional variable overloading
URL: https://github.com/apache/incubator-mxnet/pull/14438#issuecomment-479592497
@zjost thanks for the patch!
This is an automated message from the Apac
piyushghai commented on issue #14611: [WIP] Brings CI GPU images to CUDA v10.0
with latest cuDNN
URL: https://github.com/apache/incubator-mxnet/pull/14611#issuecomment-479588349
Thanks for your contributions @perdasilva.
@mxnet-label-bot Add [CUDA, CI, pr-work-in-progress]
piyushghai commented on issue #14608: [WIP]Refactor requantize
URL: https://github.com/apache/incubator-mxnet/pull/14608#issuecomment-479585800
@ZhennanQin Thanks for your contributions. Can you also look into the CI
failures ?
@mxnet-label-bot Add [pr-work-in-progress, Quantization
piyushghai commented on issue #14607: split_and_load can now handle num_ctx >
num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479585409
@mightydeveloper Thanks for your contributions.
@mxnet-label-bot Add [Gluon]
piyushghai commented on issue #14605: [Quantization] Fix quantize graph pass
URL: https://github.com/apache/incubator-mxnet/pull/14605#issuecomment-479584397
@xinyu-intel Thanks for your contributions.
@mxnet-label-bot Add [Quantization]
-
piyushghai commented on issue #14604: [MKLDNN]Add quantized relu
URL: https://github.com/apache/incubator-mxnet/pull/14604#issuecomment-479580393
@mxnet-label-bot Add [MKLDNN, pr-awaiting-review]
This is an automated message f
piyushghai commented on issue #14604: [MKLDNN]Add quantized relu
URL: https://github.com/apache/incubator-mxnet/pull/14604#issuecomment-479580308
Thanks for your contributions @huangzhiyuan.
Can you please look into the CI failures ?
piyushghai commented on issue #14269: Updated docs for R-package installation
URL: https://github.com/apache/incubator-mxnet/pull/14269#issuecomment-479577885
> @piyushghai thanks for the patch. This PR needs a rebase to the latest
master branch. Would you mind?
Done. :)
---
lanking520 merged pull request #14600: [MXNET-1379] update reshape operator
URL: https://github.com/apache/incubator-mxnet/pull/14600
This is an automated message from the Apache Git Service.
To respond to the message, please
lanking520 commented on issue #14590: Scala/Java reshape problem
URL:
https://github.com/apache/incubator-mxnet/issues/14590#issuecomment-479576973
Close this one since the problem solved
This is an automated message from the
stu1130 commented on a change in pull request #14259: Add Gluon Transformer Crop
URL: https://github.com/apache/incubator-mxnet/pull/14259#discussion_r271843416
##
File path: src/operator/image/crop-inl.h
##
@@ -0,0 +1,182 @@
+/*
+* Licensed to the Apache Software Foundatio
This is an automated email from the ASF dual-hosted git repository.
lanking 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 b482a44 [MXNET-1379] update reshape o
lanking520 closed issue #14590: Scala/Java reshape problem
URL: https://github.com/apache/incubator-mxnet/issues/14590
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitH
asmushetzel commented on issue #14501: added extraction/generation of
diagonal/triangonal matrices to linalg
URL: https://github.com/apache/incubator-mxnet/pull/14501#issuecomment-479552270
rebased.
This is an automated messa
wkcn commented on issue #14607: split_and_load can now handle num_ctx >
num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479534232
calling trainer.step(ignore_stale_grad=True) is okay : )
--
ptrendx commented on issue #14047: mxnet.base.MXNetError: Cannot find argument
'cudnn_algo_verbose'
URL:
https://github.com/apache/incubator-mxnet/issues/14047#issuecomment-479518155
Hmmm, your accuracy seems low from the start, maybe there is something wrong
with initial parameter broadc
zjost commented on issue #14438: Fixing unintentional variable overloading
URL: https://github.com/apache/incubator-mxnet/pull/14438#issuecomment-479512170
Sweet, thanks!
This is an automated message from the Apache Git Servic
mightydeveloper commented on issue #14607: split_and_load can now handle
num_ctx > num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479504030
> However, it seems that `trainer.step()` will update the weights by using
all gradients on 5 GP
david-seiler commented on a change in pull request #14579: PDF operators for
each distribution
URL: https://github.com/apache/incubator-mxnet/pull/14579#discussion_r271751995
##
File path: src/operator/random/pdf_op.cc
##
@@ -0,0 +1,292 @@
+/*
+ * Licensed to the Apache So
jrosebr1 commented on issue #14507: model links data.dmlc.ml are broken
URL:
https://github.com/apache/incubator-mxnet/issues/14507#issuecomment-479497440
Thanks @jnorwood!
This is an automated message from the Apache Git Ser
david-seiler commented on a change in pull request #14579: PDF operators for
each distribution
URL: https://github.com/apache/incubator-mxnet/pull/14579#discussion_r271751179
##
File path: src/operator/random/pdf_op.cc
##
@@ -0,0 +1,292 @@
+/*
+ * Licensed to the Apache So
david-seiler commented on a change in pull request #14579: PDF operators for
each distribution
URL: https://github.com/apache/incubator-mxnet/pull/14579#discussion_r271750980
##
File path: src/operator/random/pdf_op.h
##
@@ -0,0 +1,594 @@
+/*
+ * Licensed to the Apache Sof
david-seiler commented on a change in pull request #14579: PDF operators for
each distribution
URL: https://github.com/apache/incubator-mxnet/pull/14579#discussion_r271750744
##
File path: src/operator/random/pdf_op.h
##
@@ -0,0 +1,594 @@
+/*
+ * Licensed to the Apache Sof
david-seiler commented on a change in pull request #14579: PDF operators for
each distribution
URL: https://github.com/apache/incubator-mxnet/pull/14579#discussion_r271750062
##
File path: src/operator/random/pdf_op.cc
##
@@ -0,0 +1,292 @@
+/*
+ * Licensed to the Apache So
david-seiler commented on issue #14579: PDF operators for each distribution
URL: https://github.com/apache/incubator-mxnet/pull/14579#issuecomment-479495777
This PR now correctly targets master -- still from a branch named v1.3.x on
my side, since I couldn't see how to reset that, but that
dreadlord1984 edited a comment on issue #14471: undefined symbol:
_Z9rleDecodePK3RLEPhm
URL:
https://github.com/apache/incubator-mxnet/issues/14471#issuecomment-479490203
I meet the same error, solved by using the cocoapi code
(https://github.com/RogerChern/cocoapi). Must use this version
dreadlord1984 commented on issue #14471: undefined symbol: _Z9rleDecodePK3RLEPhm
URL:
https://github.com/apache/incubator-mxnet/issues/14471#issuecomment-479490203
I meet the same error, solved by using the cocoapi code
(https://github.com/RogerChern/cocoapi). Must use this version cocoapi
This is an automated email from the ASF dual-hosted git repository.
zhasheng 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 a51a41a Bump the publish ti
wkcn edited a comment on issue #14607: split_and_load can now handle num_ctx >
num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479478998
@mightydeveloper
Thank you for the detailed explanation!
Yes, only gradients for GPU 0, 1, 2 w
wkcn commented on issue #14607: split_and_load can now handle num_ctx >
num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479478998
@mightydeveloper
Thank you for the detailed explanation!
Yes, only gradients for GPU 0, 1, 2 will be
mightydeveloper edited a comment on issue #14607: split_and_load can now handle
num_ctx > num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479461422
@wkcn
In the example above, `losses` only contains loss terms that are forwarded
by r
mightydeveloper edited a comment on issue #14607: split_and_load can now handle
num_ctx > num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479461422
@wkcn
In the example above, `losses` only contains loss terms that are forwarded
by r
mightydeveloper commented on issue #14607: split_and_load can now handle
num_ctx > num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479461422
@wkcn
In the example above, `losses` only contains loss terms that are forwarded
by real dat
wkcn commented on issue #14610: [Feature Request] Disable lazy evaluation
URL:
https://github.com/apache/incubator-mxnet/issues/14610#issuecomment-479459608
I wrote a test case:
```python
import mxnet as mx
import time
N = 4000
a = mx.nd.zeros((N, N))
b = mx.nd.zeros((
perdasilva opened a new pull request #14611: [WIP] Brings CI GPU images to CUDA
v10.0 with latest cuDNN
URL: https://github.com/apache/incubator-mxnet/pull/14611
## Description ##
I'm trying to bring the CI GPU images up to a current version of CUDA. I
have attempted this in a previ
wkcn edited a comment on issue #14607: split_and_load can now handle num_ctx >
num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479439307
I am not sure how the weights update for the context without inpus when
backward.
-
wkcn commented on issue #14607: split_and_load can now handle num_ctx >
num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479439307
I am not sure how the weights update for the context without inpus when
backward?
mxnet-label-bot commented on issue #14610: [Feature Request] Disable lazy
evaluation
URL:
https://github.com/apache/incubator-mxnet/issues/14610#issuecomment-479429183
Hey, this is the MXNet Label Bot.
Thank you for submitting the issue! I will try and suggest some labels so
that the
wkcn opened a new issue #14610: [Feature Request] Disable lazy evaluation
URL: https://github.com/apache/incubator-mxnet/issues/14610
Hi! there.
Consider this case:
It takes T1 seconds to execute an operator `mx.nd.some_op`.
I have the code:
```python
a = mx.nd.some_op
IvyGongoogle opened a new issue #14609: The speed of `MXPredCreate` is very
slow for `mxnetv1.0.0/V100/cuda8.0/cudnn5.0/driver410`
URL: https://github.com/apache/incubator-mxnet/issues/14609
hello, I have an issue. When I use
`mxnetv1.0.0/V100/cuda10.0/cudnn7.5/driver410` to run a inferenc
mxnet-label-bot commented on issue #14609: The speed of `MXPredCreate` is very
slow for `mxnetv1.0.0/V100/cuda8.0/cudnn5.0/driver410`
URL:
https://github.com/apache/incubator-mxnet/issues/14609#issuecomment-479425807
Hey, this is the MXNet Label Bot.
Thank you for submitting the issue
lebeg commented on issue #14087: Poor performance of the libmxnet if OMP_PLACES
environment variable is present
URL:
https://github.com/apache/incubator-mxnet/issues/14087#issuecomment-479408995
Just to make a few connections between tickets:
Here was an attempt to remove bundled co
NeoZhangJianyu commented on issue #14087: Poor performance of the libmxnet if
OMP_PLACES environment variable is present
URL:
https://github.com/apache/incubator-mxnet/issues/14087#issuecomment-479399453
@stsukrov
Yes, I agree with you. I don't like to set LD_LIBRARY_PATH in fact.
Chancebair commented on issue #14161: Disable flaky test
test_operator_gpu.test_multinomial_generator
URL: https://github.com/apache/incubator-mxnet/pull/14161#issuecomment-479390262
The issue tracking this flaky test is listed in the first comment. The way
we handle flaky tests once they
Chancebair closed pull request #14161: Disable flaky test
test_operator_gpu.test_multinomial_generator
URL: https://github.com/apache/incubator-mxnet/pull/14161
This is an automated message from the Apache Git Service.
To re
ZhennanQin opened a new pull request #14608: [WIP]Refactor requantize
URL: https://github.com/apache/incubator-mxnet/pull/14608
## Description ##
Currently, requantize is too slow to use. This PR is to refactor its MKLDNN
implementation to make it usable.
## Checklist ##
### Es
stsukrov commented on issue #14087: Poor performance of the libmxnet if
OMP_PLACES environment variable is present
URL:
https://github.com/apache/incubator-mxnet/issues/14087#issuecomment-479387571
@NeoZhangJianyu
Thanks for providing work around recommendations.
Was actual
stsukrov edited a comment on issue #14087: Poor performance of the libmxnet if
OMP_PLACES environment variable is present
URL:
https://github.com/apache/incubator-mxnet/issues/14087#issuecomment-461733006
@pengzhao-intel
Thanks, Patric. Appreciate your help during the vacation.
NeoZhangJianyu edited a comment on issue #14087: Poor performance of the
libmxnet if OMP_PLACES environment variable is present
URL:
https://github.com/apache/incubator-mxnet/issues/14087#issuecomment-479381528
@sandeep-krishnamurthy
"This issue is caused by 2 conflicting OMP runtimes"
NeoZhangJianyu edited a comment on issue #14087: Poor performance of the
libmxnet if OMP_PLACES environment variable is present
URL:
https://github.com/apache/incubator-mxnet/issues/14087#issuecomment-479381528
@sandeep-krishnamurthy
"This issue is caused by 2 conflicting OMP runtimes"
NeoZhangJianyu commented on issue #14087: Poor performance of the libmxnet if
OMP_PLACES environment variable is present
URL:
https://github.com/apache/incubator-mxnet/issues/14087#issuecomment-479381528
@Sandeep Krishnamurthy
"This issue is caused by 2 conflicting OMP runtimes" as ment
mightydeveloper commented on issue #14607: split_and_load can now handle
num_ctx > num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607#issuecomment-479378305
Just an FYI, I made a misleading comment on the code previously, so I fixed
it.
--
david-seiler commented on issue #14579: PDF operators for each distribution
URL: https://github.com/apache/incubator-mxnet/pull/14579#issuecomment-479376868
> @david-seiler Seems like you want to add a particular feature to a
previous version of MXNet.
I started with 1.3.x because it
mightydeveloper opened a new pull request #14607: split_and_load can now handle
num_ctx > num_data. Github Issue #13909
URL: https://github.com/apache/incubator-mxnet/pull/14607
## Description ##
Handles Issue #13909
when last batch size is smaller than the number of contexts,
the
sxjscience edited a comment on issue #14073: Fp16 support for layernorm
URL:
https://github.com/apache/incubator-mxnet/issues/14073#issuecomment-478921540
Current I propose to solve the issue following these two steps:
- [ ] Add a new 'reduer_dtype' flag to the LayerNorm operator. Wh
This is an automated email from the ASF dual-hosted git repository.
zhasheng 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 c02d810 Bump the publish ti
101 - 176 of 176 matches
Mail list logo