[GitHub] [incubator-mxnet] perdasilva commented on issue #14336: CI Changes for Codified Windows AMIs

perdasilva commented on issue #14336: CI Changes for Codified Windows AMIs URL: https://github.com/apache/incubator-mxnet/pull/14336#issuecomment-470837197 Need to disabling test_operator_gpu.test_bulking (#14368). It's a new test, that hasn't been through the windows pipeline, which is

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #14368: test_operator_gpu.test_bulking failing on windows

mxnet-label-bot commented on issue #14368: test_operator_gpu.test_bulking failing on windows URL: https://github.com/apache/incubator-mxnet/issues/14368#issuecomment-470836850 Hey, this is the MXNet Label Bot. Thank you for submitting the issue! I will try and suggest some labels so

[GitHub] [incubator-mxnet] perdasilva opened a new issue #14368: test_operator_gpu.test_bulking failing on windows

perdasilva opened a new issue #14368: test_operator_gpu.test_bulking failing on windows URL: https://github.com/apache/incubator-mxnet/issues/14368 ## Description When re-enabling the windows pipeline (#14336), I found that test_operator_gpu.test_bulking is [not

[GitHub] [incubator-mxnet] seujung commented on issue #13966: upload example code for snail with gluon

seujung commented on issue #13966: upload example code for snail with gluon URL: https://github.com/apache/incubator-mxnet/pull/13966#issuecomment-470834456 For the push issue, I close the Pull Request. I make a new pr #14367

[GitHub] [incubator-mxnet] seujung opened a new pull request #14367: upload snail for gluon

seujung opened a new pull request #14367: upload snail for gluon URL: https://github.com/apache/incubator-mxnet/pull/14367 ## Description ## upload example for snail code SNAIL(A Simple Neural Attentive Meta-Learniner) is a kind of meta-learning algorithm. Train network using

[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.

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 f9da1ea Bump the publish

[GitHub] [incubator-mxnet] seujung closed pull request #13966: upload example code for snail with gluon

seujung closed pull request #13966: upload example code for snail with gluon URL: https://github.com/apache/incubator-mxnet/pull/13966 This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [incubator-mxnet] eric-haibin-lin opened a new issue #14366: Flaky test test_with_random_seed

eric-haibin-lin opened a new issue #14366: Flaky test test_with_random_seed URL: https://github.com/apache/incubator-mxnet/issues/14366 Appeared in https://github.com/apache/incubator-mxnet/pull/14321 See logs at -

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #14366: Flaky test test_with_random_seed

mxnet-label-bot commented on issue #14366: Flaky test test_with_random_seed URL: https://github.com/apache/incubator-mxnet/issues/14366#issuecomment-470825995 Hey, this is the MXNet Label Bot. Thank you for submitting the issue! I will try and suggest some labels so that the

[GitHub] [incubator-mxnet] Ian8888 commented on issue #14333: Predict of mx.mod.Module api problem

Ian commented on issue #14333: Predict of mx.mod.Module api problem URL: https://github.com/apache/incubator-mxnet/issues/14333#issuecomment-470824608 > Yes, it's normal. You should add a argmax layer on the top of the network to get the label Thanks, your suggest. I can get

[GitHub] [incubator-mxnet] eric-haibin-lin commented on issue #14077: FP16 support for contrib.adamw

eric-haibin-lin commented on issue #14077: FP16 support for contrib.adamw URL: https://github.com/apache/incubator-mxnet/issues/14077#issuecomment-470823605 Added in https://github.com/apache/incubator-mxnet/pull/14171 This

[GitHub] [incubator-mxnet] eric-haibin-lin closed issue #14077: FP16 support for contrib.adamw

eric-haibin-lin closed issue #14077: FP16 support for contrib.adamw URL: https://github.com/apache/incubator-mxnet/issues/14077 This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] [incubator-mxnet] eric-haibin-lin closed issue #14257: Flaky cpp test: dropout gpu perf

eric-haibin-lin closed issue #14257: Flaky cpp test: dropout gpu perf URL: https://github.com/apache/incubator-mxnet/issues/14257 This is an automated message from the Apache Git Service. To respond to the message, please

[GitHub] [incubator-mxnet] eric-haibin-lin commented on issue #14257: Flaky cpp test: dropout gpu perf

eric-haibin-lin commented on issue #14257: Flaky cpp test: dropout gpu perf URL: https://github.com/apache/incubator-mxnet/issues/14257#issuecomment-470823284 #14278 fixes the stream used in the test. feel free to reopen if this issue happens again

[GitHub] [incubator-mxnet] eric-haibin-lin commented on issue #14072: Fp16 support for softmax

eric-haibin-lin commented on issue #14072: Fp16 support for softmax URL: https://github.com/apache/incubator-mxnet/issues/14072#issuecomment-470822886 fixed in https://github.com/apache/incubator-mxnet/pull/14098 This is an

[GitHub] [incubator-mxnet] eric-haibin-lin closed issue #14072: Fp16 support for softmax

eric-haibin-lin closed issue #14072: Fp16 support for softmax URL: https://github.com/apache/incubator-mxnet/issues/14072 This is an automated message from the Apache Git Service. To respond to the message, please log on to

[incubator-mxnet] branch master updated: add pos_weight for SigmoidBinaryCrossEntropyLoss (#13612)

This is an automated email from the ASF dual-hosted git repository. wkcn 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 ce9e3cf add pos_weight for

[GitHub] [incubator-mxnet] wkcn commented on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss

wkcn commented on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss URL: https://github.com/apache/incubator-mxnet/pull/13612#issuecomment-470817331 The PR has been merged. Thanks for your contribution: ) This

[GitHub] [incubator-mxnet] wkcn merged pull request #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss

wkcn merged pull request #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss URL: https://github.com/apache/incubator-mxnet/pull/13612 This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-mxnet] wkcn commented on issue #14344: [WIP]support multi-theading for Gluon BlockScope

wkcn commented on issue #14344: [WIP]support multi-theading for Gluon BlockScope URL: https://github.com/apache/incubator-mxnet/pull/14344#issuecomment-470816664 @eric-haibin-lin Sorry that it is difficult for me to fix the issue. I found there are many variables to write value, and do

[GitHub] [incubator-mxnet] vandanavk commented on issue #14365: Fix memory leak for size-zero ndarray

vandanavk commented on issue #14365: Fix memory leak for size-zero ndarray URL: https://github.com/apache/incubator-mxnet/pull/14365#issuecomment-470813818 @yuxihu Please add "Fixes https://github.com/apache/incubator-mxnet/issues/14358; as well, so that #14358 is also closed when this PR

[GitHub] [incubator-mxnet] vandanavk commented on issue #14363: Support multi-threading for Custom Operator

vandanavk commented on issue #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#issuecomment-470813575 @mxnet-label-bot add [Operator, pr-awaiting-review] This is an

[GitHub] [incubator-mxnet] wkcn commented on a change in pull request #14363: Support multi-threading for Custom Operator

wkcn commented on a change in pull request #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#discussion_r263666113 ## File path: src/operator/custom/custom-inl.h ## @@ -139,38 +142,52 @@ class CustomOperator {

[GitHub] [incubator-mxnet] wkcn commented on a change in pull request #14363: Support multi-threading for Custom Operator

wkcn commented on a change in pull request #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#discussion_r263669472 ## File path: src/operator/custom/custom-inl.h ## @@ -129,6 +129,9 @@ class CustomOperator {

[GitHub] [incubator-mxnet] wkcn commented on a change in pull request #14363: Support multi-threading for Custom Operator

wkcn commented on a change in pull request #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#discussion_r263669472 ## File path: src/operator/custom/custom-inl.h ## @@ -129,6 +129,9 @@ class CustomOperator {

[GitHub] [incubator-mxnet] TaoLv commented on issue #11742: [MXNET-623] Fixing an integer overflow bug in large NDArray

TaoLv commented on issue #11742: [MXNET-623] Fixing an integer overflow bug in large NDArray URL: https://github.com/apache/incubator-mxnet/pull/11742#issuecomment-470809366 @apeforest Yes, I know the script is there. I'm just wondering if it's actually executed nightly and where can I

[GitHub] [incubator-mxnet] eric-haibin-lin commented on issue #14344: [WIP]support multi-theading for Gluon BlockScope

eric-haibin-lin commented on issue #14344: [WIP]support multi-theading for Gluon BlockScope URL: https://github.com/apache/incubator-mxnet/pull/14344#issuecomment-470806949 Look forward to the final version. I previously hit similar issues and found many thread_local variables, including

[GitHub] [incubator-mxnet] wkcn commented on a change in pull request #14363: Support multi-threading for Custom Operator

wkcn commented on a change in pull request #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#discussion_r263666113 ## File path: src/operator/custom/custom-inl.h ## @@ -139,38 +142,52 @@ class CustomOperator {

[GitHub] [incubator-mxnet] stereomatchingkiss commented on issue #14364: Enable MKL_USE_ILP64 cause a lot of compile time errors on windws

stereomatchingkiss commented on issue #14364: Enable MKL_USE_ILP64 cause a lot of compile time errors on windws URL: https://github.com/apache/incubator-mxnet/issues/14364#issuecomment-470804216 I "fix" those compile time errors, but this do not solve the issue, because this time it pop

[GitHub] [incubator-mxnet] eric-haibin-lin commented on a change in pull request #14363: Support multi-threading for Custom Operator

eric-haibin-lin commented on a change in pull request #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#discussion_r263663529 ## File path: src/operator/custom/custom-inl.h ## @@ -129,6 +129,9 @@ class

[GitHub] [incubator-mxnet] eric-haibin-lin commented on a change in pull request #14363: Support multi-threading for Custom Operator

eric-haibin-lin commented on a change in pull request #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#discussion_r263663299 ## File path: src/operator/custom/custom-inl.h ## @@ -139,38 +142,52 @@ class

[GitHub] [incubator-mxnet] yuxihu edited a comment on issue #14365: Fix memory leak for size-zero ndarray

yuxihu edited a comment on issue #14365: Fix memory leak for size-zero ndarray URL: https://github.com/apache/incubator-mxnet/pull/14365#issuecomment-470797391 @mxnet-label-bot update [pr-work-in-progress] This is an

[GitHub] [incubator-mxnet] yuxihu edited a comment on issue #14365: Fix memory leak for size-zero ndarray

yuxihu edited a comment on issue #14365: Fix memory leak for size-zero ndarray URL: https://github.com/apache/incubator-mxnet/pull/14365#issuecomment-470797391 @mxnet-label-bot add [pr-work-in-progress] This is an automated

[GitHub] [incubator-mxnet] TaoLv merged pull request #14128: MKLDNN based Quantized FullyConnected Operator and its fusion

TaoLv merged pull request #14128: MKLDNN based Quantized FullyConnected Operator and its fusion URL: https://github.com/apache/incubator-mxnet/pull/14128 This is an automated message from the Apache Git Service. To respond

[incubator-mxnet] branch master updated: MKLDNN based Quantized FullyConnected Operator and its fusion (#14128)

This is an automated email from the ASF dual-hosted git repository. taolv 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 8668db7 MKLDNN based Quantized

[GitHub] [incubator-mxnet] TaoLv commented on issue #14128: MKLDNN based Quantized FullyConnected Operator and its fusion

TaoLv commented on issue #14128: MKLDNN based Quantized FullyConnected Operator and its fusion URL: https://github.com/apache/incubator-mxnet/pull/14128#issuecomment-470801817 Thank you for the contribution @ciyongch. Merging now.

[GitHub] [incubator-mxnet] stereomatchingkiss edited a comment on issue #14343: Cannot generate op.h when build with mkl

stereomatchingkiss edited a comment on issue #14343: Cannot generate op.h when build with mkl URL: https://github.com/apache/incubator-mxnet/issues/14343#issuecomment-470794886 I guess I found the issue, you need to enable MKL_USE_ILP64(I am surprise it can compile before I actiave it,

[GitHub] [incubator-mxnet] stereomatchingkiss edited a comment on issue #14343: Cannot generate op.h when build with mkl

stereomatchingkiss edited a comment on issue #14343: Cannot generate op.h when build with mkl URL: https://github.com/apache/incubator-mxnet/issues/14343#issuecomment-470794886 I guess I found the issue, you need to enable MKL_USE_ILP64(I am surprise it can compile before I actiave it,

[GitHub] [incubator-mxnet] yuxihu commented on issue #14365: Fix memory leak for size-zero ndarray

yuxihu commented on issue #14365: Fix memory leak for size-zero ndarray URL: https://github.com/apache/incubator-mxnet/pull/14365#issuecomment-470797391 @yifeim @apeforest @anirudh2290 please review. @mxnet-label-bot add [pr-awaiting-review]

[GitHub] [incubator-mxnet] yuxihu opened a new pull request #14365: Fix memory leak for size-zero ndarray

yuxihu opened a new pull request #14365: Fix memory leak for size-zero ndarray URL: https://github.com/apache/incubator-mxnet/pull/14365 Fix #13951 #14358 For size-zero ndarray (e.g. mx.nd.array([]), mx.nd.ones(0)), the storage handle size is 0. Currently we only free handles which

[GitHub] [incubator-mxnet] szha commented on issue #14262: Fix NaN value comparisons in relu, max and min ops

szha commented on issue #14262: Fix NaN value comparisons in relu, max and min ops URL: https://github.com/apache/incubator-mxnet/pull/14262#issuecomment-470796792 @anirudhacharya one last thing, could you measure the performance before and after this change? This change is nonetheless

[GitHub] [incubator-mxnet] wkcn edited a comment on issue #14363: Support multi-threading for Custom Operator

wkcn edited a comment on issue #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#issuecomment-470794994 I will check whehter it is correct to use the lock. it seems that there is no problem.

[GitHub] [incubator-mxnet] wkcn edited a comment on issue #14363: Support multi-threading for Custom Operator

wkcn edited a comment on issue #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#issuecomment-470794994 I will check whehter it is correct to use the lock. Seems that there is no problem.

[GitHub] [incubator-mxnet] yuxihu edited a comment on issue #14358: Memory builds up when creating size-zero NDArray in a loop

yuxihu edited a comment on issue #14358: Memory builds up when creating size-zero NDArray in a loop URL: https://github.com/apache/incubator-mxnet/issues/14358#issuecomment-470796343 @mxnet-label-bot add [Backend, Memory]

[GitHub] [incubator-mxnet] yuxihu commented on issue #14358: Memory builds up when creating size-zero NDArray in a loop

yuxihu commented on issue #14358: Memory builds up when creating size-zero NDArray in a loop URL: https://github.com/apache/incubator-mxnet/issues/14358#issuecomment-470796343 @mxnet-label-bot add [Backend] This is an

[GitHub] [incubator-mxnet] stereomatchingkiss opened a new issue #14364: Enable MKL_USE_ILP64 cause a lot of compile time errors on windws

stereomatchingkiss opened a new issue #14364: Enable MKL_USE_ILP64 cause a lot of compile time errors on windws URL: https://github.com/apache/incubator-mxnet/issues/14364 - os : windows 10 64bits - compiler : vc2015 64bits update 3 - mxnet : 1.3.1 As the title mentioned,

[GitHub] [incubator-mxnet] mxnet-label-bot commented on issue #14364: Enable MKL_USE_ILP64 cause a lot of compile time errors on windws

mxnet-label-bot commented on issue #14364: Enable MKL_USE_ILP64 cause a lot of compile time errors on windws URL: https://github.com/apache/incubator-mxnet/issues/14364#issuecomment-470796188 Hey, this is the MXNet Label Bot. Thank you for submitting the issue! I will try and suggest

[GitHub] [incubator-mxnet] anirudh2290 commented on issue #13265: MXNet C++Interface reasoning leads to CPU memory leak

anirudh2290 commented on issue #13265: MXNet C++Interface reasoning leads to CPU memory leak URL: https://github.com/apache/incubator-mxnet/issues/13265#issuecomment-470795561 Thanks for the great discussion! @JohnLee168 and @coolwebgo are you guys still seeing memory leak ? Can you

[GitHub] [incubator-mxnet] wkcn commented on issue #14363: Support multi-threading for Custom Operator

wkcn commented on issue #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363#issuecomment-470794994 I will check whehter it is correct to use the lock. This is an

[GitHub] [incubator-mxnet] stereomatchingkiss edited a comment on issue #14343: Cannot generate op.h when build with mkl

stereomatchingkiss edited a comment on issue #14343: Cannot generate op.h when build with mkl URL: https://github.com/apache/incubator-mxnet/issues/14343#issuecomment-470794886 I guess I found the issue, you need to enable MKL_USE_ILP64(I am surprise it can compile before I actiave it,

[GitHub] [incubator-mxnet] stereomatchingkiss commented on issue #14343: Cannot generate op.h when build with mkl

stereomatchingkiss commented on issue #14343: Cannot generate op.h when build with mkl URL: https://github.com/apache/incubator-mxnet/issues/14343#issuecomment-470794886 I guess I found the issue, you need to enable MKL_USE_ILP64(I am surprise it can compile even my pc only install MKL

[GitHub] [incubator-mxnet] wkcn opened a new pull request #14363: Support multi-threading for Custom Operator

wkcn opened a new pull request #14363: Support multi-threading for Custom Operator URL: https://github.com/apache/incubator-mxnet/pull/14363 ## Description ## Hi! I found that there was single-thread to execute custom operator. It drops the performance. Is it a better way to support

[GitHub] [incubator-mxnet] anirudh2290 commented on issue #12128: NaiveEngine is not threadsafe and crashes when called from multiple threads.

anirudh2290 commented on issue #12128: NaiveEngine is not threadsafe and crashes when called from multiple threads. URL: https://github.com/apache/incubator-mxnet/issues/12128#issuecomment-470792609 Closing since non-issue. @anirudhacharya can you please also check gluon and open a new

[GitHub] [incubator-mxnet] anirudh2290 closed issue #12128: NaiveEngine is not threadsafe and crashes when called from multiple threads.

anirudh2290 closed issue #12128: NaiveEngine is not threadsafe and crashes when called from multiple threads. URL: https://github.com/apache/incubator-mxnet/issues/12128 This is an automated message from the Apache Git

[GitHub] [incubator-mxnet] anirudh2290 closed issue #9166: Segfault on ndarray with negative dimension i.e. mxnet.nx.zeros((-1, ))

anirudh2290 closed issue #9166: Segfault on ndarray with negative dimension i.e. mxnet.nx.zeros((-1,)) URL: https://github.com/apache/incubator-mxnet/issues/9166 This is an automated message from the Apache Git Service. To

[incubator-mxnet] branch master updated: add exception (#14362)

This is an automated email from the ASF dual-hosted git repository. anirudh2290 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 30b1cbc add exception (#14362)

[GitHub] [incubator-mxnet] anirudh2290 merged pull request #14362: Add proper exception message for negative shape in array creation routines

anirudh2290 merged pull request #14362: Add proper exception message for negative shape in array creation routines URL: https://github.com/apache/incubator-mxnet/pull/14362 This is an automated message from the Apache Git

[GitHub] [incubator-mxnet] e8035669 commented on issue #14184: Cannot infer RNN symbol with mode=RNNMode::kRnn_tanh

e8035669 commented on issue #14184: Cannot infer RNN symbol with mode=RNNMode::kRnn_tanh URL: https://github.com/apache/incubator-mxnet/issues/14184#issuecomment-470791193 Hi @leleamol Thanks for your reply. I have tried to pass a dummy state_cell symbol into it, But it will always

[GitHub] [incubator-mxnet] apeforest commented on issue #11742: [MXNET-623] Fixing an integer overflow bug in large NDArray

apeforest commented on issue #11742: [MXNET-623] Fixing an integer overflow bug in large NDArray URL: https://github.com/apache/incubator-mxnet/pull/11742#issuecomment-470788806 @TaoLv It's in tests/nightly/test_large_array.py

[GitHub] [incubator-mxnet] zhreshold commented on issue #14058: add backgroud class in box_nms

zhreshold commented on issue #14058: add backgroud class in box_nms URL: https://github.com/apache/incubator-mxnet/pull/14058#issuecomment-470787355 @arcadiaphy Please resolve the conflict and it's good to go.  This is an

[GitHub] [incubator-mxnet] zhreshold merged pull request #14352: Optimize NMS part 2

zhreshold merged pull request #14352: Optimize NMS part 2 URL: https://github.com/apache/incubator-mxnet/pull/14352 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub

[incubator-mxnet] branch master updated: Optimize NMS part 2 (#14352)

This is an automated email from the ASF dual-hosted git repository. zhreshold 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 838e256 Optimize NMS part 2

[GitHub] [incubator-mxnet] TaoLv commented on issue #14202: [Doc] Start the tutorials for MKL-DNN backend

TaoLv commented on issue #14202: [Doc] Start the tutorials for MKL-DNN backend URL: https://github.com/apache/incubator-mxnet/pull/14202#issuecomment-470785347 Any comments? @aaronmarkham This is an automated message from the

[GitHub] [incubator-mxnet] wkcn commented on issue #13199: Namescope is None when hybridize in multi-threading environment. AttributeError: 'NoneType' object has no attribute '__exit__'

wkcn commented on issue #13199: Namescope is None when hybridize in multi-threading environment. AttributeError: 'NoneType' object has no attribute '__exit__' URL: https://github.com/apache/incubator-mxnet/issues/13199#issuecomment-470784948 I see. There is only one thread to execute

[GitHub] [incubator-mxnet] TaoLv commented on issue #11742: [MXNET-623] Fixing an integer overflow bug in large NDArray

TaoLv commented on issue #11742: [MXNET-623] Fixing an integer overflow bug in large NDArray URL: https://github.com/apache/incubator-mxnet/pull/11742#issuecomment-470784123 @apeforest @marcoabreu Can you point me to the tested history of the new test_large_ndarray.py script? Cannot find

[GitHub] [incubator-mxnet] AmigoCDT commented on issue #14229: About the weighted softmax, the forward result is too small

AmigoCDT commented on issue #14229: About the weighted softmax, the forward result is too small URL: https://github.com/apache/incubator-mxnet/issues/14229#issuecomment-470780197 @piyushghai I think the error is in backward() or gradient passing. I find few user rewrite

[GitHub] [incubator-mxnet] AmigoCDT edited a comment on issue #14229: About the weighted softmax, the forward result is too small

AmigoCDT edited a comment on issue #14229: About the weighted softmax, the forward result is too small URL: https://github.com/apache/incubator-mxnet/issues/14229#issuecomment-470778540 @piyushghai yes, I use another similar customOP (Focal loss), it works well. So the trouble is in

[GitHub] [incubator-mxnet] AmigoCDT commented on issue #14229: About the weighted softmax, the forward result is too small

AmigoCDT commented on issue #14229: About the weighted softmax, the forward result is too small URL: https://github.com/apache/incubator-mxnet/issues/14229#issuecomment-470778540 @piyushghai yes, I use another similar customOP (Focal loss), it works well. So the trouble is in my code.

[GitHub] [incubator-mxnet] wkcn edited a comment on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss

wkcn edited a comment on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss URL: https://github.com/apache/incubator-mxnet/pull/13612#issuecomment-470773887 @eureka7mt I think we may pass Symbol into SigmoidBinaryCrossEntropyLoss. Symbol will not broadcast_mul automatically in

[GitHub] [incubator-mxnet] wkcn edited a comment on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss

wkcn edited a comment on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss URL: https://github.com/apache/incubator-mxnet/pull/13612#issuecomment-470773887 @eureka7mt I think we may pass Symbol into SigmoidBinaryCrossEntropyLoss. Symbol will not broadcast_mul automatically in

[GitHub] [incubator-mxnet] wkcn commented on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss

wkcn commented on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss URL: https://github.com/apache/incubator-mxnet/pull/13612#issuecomment-470773887 @eureka7mt I think we may pass Symbol into SigmoidBinaryCrossEntropyLoss.

[GitHub] [incubator-mxnet] eureka7mt commented on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss

eureka7mt commented on issue #13612: add pos_weight for SigmoidBinaryCrossEntropyLoss URL: https://github.com/apache/incubator-mxnet/pull/13612#issuecomment-470771170 Maybe the broadcast_mul isn't necessary.I think that a NDArray * a NDArray will do broadcast_mul automatically.

[GitHub] [incubator-mxnet] leleamol edited a comment on issue #14184: Cannot infer RNN symbol with mode=RNNMode::kRnn_tanh

leleamol edited a comment on issue #14184: Cannot infer RNN symbol with mode=RNNMode::kRnn_tanh URL: https://github.com/apache/incubator-mxnet/issues/14184#issuecomment-470770974 Hi @e8035669 The "op.h" file in cpp-package is automatically generated. It contains the function

[GitHub] [incubator-mxnet] leleamol commented on issue #14184: Cannot infer RNN symbol with mode=RNNMode::kRnn_tanh

leleamol commented on issue #14184: Cannot infer RNN symbol with mode=RNNMode::kRnn_tanh URL: https://github.com/apache/incubator-mxnet/issues/14184#issuecomment-470770974 Hi @e8035669 The "op.h" file in cpp-package is automatically generated. It contains the function declaration

[GitHub] [incubator-mxnet] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r263638954 ## File path: example/distributed_training-horovod/README.md ## @@ -0,0 +1,207 @@ + + + + + +

[GitHub] [incubator-mxnet] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r263639018 ## File path: example/distributed_training-horovod/mxnet_imagenet_resnet50.py ## @@ -0,0

[GitHub] [incubator-mxnet] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r263638829 ## File path: example/distributed_training-horovod/mxnet_mnist.py ## @@ -0,0 +1,142 @@ +#

[GitHub] [incubator-mxnet] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r263638882 ## File path: example/distributed_training-horovod/README.md ## @@ -0,0 +1,207 @@ + + + + + +

[GitHub] [incubator-mxnet] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r263638782 ## File path: example/distributed_training-horovod/mxnet_imagenet_resnet50.py ## @@ -0,0

[GitHub] [incubator-mxnet] apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod

apeforest commented on a change in pull request #14286: Add examples of running MXNet with Horovod URL: https://github.com/apache/incubator-mxnet/pull/14286#discussion_r263638720 ## File path: example/distributed_training-horovod/README.md ## @@ -0,0 +1,207 @@ + + + + + +

[GitHub] [incubator-mxnet] yifeim commented on issue #13951: Gluon RNN memory leaks with extra variables

yifeim commented on issue #13951: Gluon RNN memory leaks with extra variables URL: https://github.com/apache/incubator-mxnet/issues/13951#issuecomment-470767282 Very interesting. Thanks a lot for the insights! This is an

[GitHub] [incubator-mxnet] wkcn commented on issue #13199: Namescope is None when hybridize in multi-threading environment. AttributeError: 'NoneType' object has no attribute '__exit__'

wkcn commented on issue #13199: Namescope is None when hybridize in multi-threading environment. AttributeError: 'NoneType' object has no attribute '__exit__' URL: https://github.com/apache/incubator-mxnet/issues/13199#issuecomment-470766959 @kohillyang Hi! Could you please provide an

[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.

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 c5869c5 Bump the publish

[GitHub] [incubator-mxnet] leleamol commented on issue #13870: [C++] Linking static library (error or bug?) to load python trained model

leleamol commented on issue #13870: [C++] Linking static library (error or bug?) to load python trained model URL: https://github.com/apache/incubator-mxnet/issues/13870#issuecomment-470765599 @hengdos I was trying to reproduce this issue on Mac and Ubuntu. But I was not able to invoke

[GitHub] [incubator-mxnet] anirudhacharya commented on issue #14317: std::threads spawned by mxnet must catch all exceptions, otherwise the whole application will terminate

anirudhacharya commented on issue #14317: std::threads spawned by mxnet must catch all exceptions, otherwise the whole application will terminate URL: https://github.com/apache/incubator-mxnet/issues/14317#issuecomment-470764520 @mxnet-label-bot add [Thread Safety]

[GitHub] [incubator-mxnet] piyushghai commented on issue #14345: Output inconsistency between mxnet model and converted onnx model

piyushghai commented on issue #14345: Output inconsistency between mxnet model and converted onnx model URL: https://github.com/apache/incubator-mxnet/issues/14345#issuecomment-470760494 @XinyuDu I downloaded your model, and given the code snippet, I was not able to convert it to ONNX.

[GitHub] [incubator-mxnet] anirudhacharya edited a comment on issue #14262: Fix NaN value comparisons in relu, max and min ops

anirudhacharya edited a comment on issue #14262: Fix NaN value comparisons in relu, max and min ops URL: https://github.com/apache/incubator-mxnet/pull/14262#issuecomment-470760203 I modified relu grad to also propagate `NaN` values. As discussed above, since `NaN` does not exist in the

[GitHub] [incubator-mxnet] anirudhacharya commented on issue #14262: Fix NaN value comparisons in relu, max and min ops

anirudhacharya commented on issue #14262: Fix NaN value comparisons in relu, max and min ops URL: https://github.com/apache/incubator-mxnet/pull/14262#issuecomment-470760203 I modilied relu grad to also propagate `NaN` values. As discussed above, since `NaN` does not exist in the domain

[GitHub] [incubator-mxnet] anirudhacharya opened a new pull request #14362: Add proper exception message for negative shape in array creation routines

anirudhacharya opened a new pull request #14362: Add proper exception message for negative shape in array creation routines URL: https://github.com/apache/incubator-mxnet/pull/14362 ## Description ## Fixes https://github.com/apache/incubator-mxnet/issues/9166 Addresses @apeforest 's

[GitHub] [incubator-mxnet] wkcn commented on issue #14154: Improving the run time of the tests which use assert_almost_equal OR check_numeric_gradient functions

wkcn commented on issue #14154: Improving the run time of the tests which use assert_almost_equal OR check_numeric_gradient functions URL: https://github.com/apache/incubator-mxnet/pull/14154#issuecomment-470745545 Is it available to rebase this PR? It is too many changed files showed in

[GitHub] [incubator-mxnet] lanking520 commented on issue #14174: CI fails for unix cpu Static build CPU 14.04 Scala

lanking520 commented on issue #14174: CI fails for unix cpu Static build CPU 14.04 Scala URL: https://github.com/apache/incubator-mxnet/issues/14174#issuecomment-470745583 @piyushghai Well, it both happened. This is an

[GitHub] [incubator-mxnet] stereomatchingkiss edited a comment on issue #14343: Cannot generate op.h when build with mkl

stereomatchingkiss edited a comment on issue #14343: Cannot generate op.h when build with mkl URL: https://github.com/apache/incubator-mxnet/issues/14343#issuecomment-470584244 > If you use opencv to read the image, you might need the release folder of opencv in your PATH. Yes, I

[GitHub] [incubator-mxnet] vandanavk commented on issue #14359: Speedup _contrib_index_copy

vandanavk commented on issue #14359: Speedup _contrib_index_copy URL: https://github.com/apache/incubator-mxnet/pull/14359#issuecomment-470744574 @mxnet-label-bot add [Operator, pr-awaiting-review] This is an automated

[GitHub] [incubator-mxnet] wkcn commented on issue #14321: moveaxis operator now accepts negative indices and sequence of ints as well.

wkcn commented on issue #14321: moveaxis operator now accepts negative indices and sequence of ints as well. URL: https://github.com/apache/incubator-mxnet/pull/14321#issuecomment-470744294 Sorry that I wrote a typo, which should be retrigger.

[GitHub] [incubator-mxnet] reminisce opened a new pull request #14361: [numpy] Add a global switch to turn on/off numpy compatibility

reminisce opened a new pull request #14361: [numpy] Add a global switch to turn on/off numpy compatibility URL: https://github.com/apache/incubator-mxnet/pull/14361 This PR is implemented based upon the discussion with @eric-haibin-lin. It provides two APIs and NumPy-compatible NDArray

[GitHub] [incubator-mxnet] lanking520 commented on issue #12710: Subprocess Deadlock with mxnet-mkl

lanking520 commented on issue #12710: Subprocess Deadlock with mxnet-mkl URL: https://github.com/apache/incubator-mxnet/issues/12710#issuecomment-470741284 Close this issue for now. Please feel free to reopen it if you are facing more problems with it.

[GitHub] [incubator-mxnet] lanking520 commented on issue #13800: [Nightly test] v1.3.x failing with missing cmake

lanking520 commented on issue #13800: [Nightly test] v1.3.x failing with missing cmake URL: https://github.com/apache/incubator-mxnet/issues/13800#issuecomment-470741102 Close this issue for now. Please feel free to reopen it if you are facing more problems with it.

[GitHub] [incubator-mxnet] lanking520 closed issue #12710: Subprocess Deadlock with mxnet-mkl

lanking520 closed issue #12710: Subprocess Deadlock with mxnet-mkl URL: https://github.com/apache/incubator-mxnet/issues/12710 This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] [incubator-mxnet] lanking520 closed issue #13800: [Nightly test] v1.3.x failing with missing cmake

lanking520 closed issue #13800: [Nightly test] v1.3.x failing with missing cmake URL: https://github.com/apache/incubator-mxnet/issues/13800 This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-mxnet] louisfeng commented on issue #12502: [NGRAPH] MXNet - nGraph initial integration

louisfeng commented on issue #12502: [NGRAPH] MXNet - nGraph initial integration URL: https://github.com/apache/incubator-mxnet/pull/12502#issuecomment-470741022 @szha Our bridge code can be treated as 3rd party submodule and is publicly open source:

  1   2   3   >