[jira] [Commented] (MXNET-33) mxnet mkl-dnn crashes for Max pooling with convention "full"
[ https://issues.apache.org/jira/browse/MXNET-33?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16392127#comment-16392127 ] ASF GitHub Bot commented on MXNET-33: - marcoabreu commented on issue #10021: [MXNET-33] Mkldnn pooling convention crash URL: https://github.com/apache/incubator-mxnet/pull/10021#issuecomment-371665164 Feel free to develop and add an appropriate unit test yourself. Please don't hesitate to reach out to us in case you need assistance. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > mxnet mkl-dnn crashes for Max pooling with convention "full" > - > > Key: MXNET-33 > URL: https://issues.apache.org/jira/browse/MXNET-33 > Project: Apache MXNet > Issue Type: Bug >Reporter: Ashok Emani >Assignee: Ashok Emani >Priority: Minor > > A crash is seen when running SSD-VGG16 300 VOC , with mxnet mkldnn build, for > the following op. > > "op": "Pooling", > "attrs": { > "kernel": "(2, 2)", > "pool_type": "max", > "pooling_convention": "full", > "stride": "(2, 2)" > }, > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Commented] (MXNET-33) mxnet mkl-dnn crashes for Max pooling with convention "full"
[ https://issues.apache.org/jira/browse/MXNET-33?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16391964#comment-16391964 ] ASF GitHub Bot commented on MXNET-33: - ashokei commented on issue #10021: [WIP] [MXNET-33] Mkldnn pooling convention crash URL: https://github.com/apache/incubator-mxnet/pull/10021#issuecomment-371638503 @marcoabreu @TaoLv @zheng-da Unit-tests currently do not seem to test backward/training mode for pooling very well. Below is the test for reproducing this crash. ` cd example/image-classification/symbols/ ln -s ../../ssd/symbol/vgg16_reduced.py vgg16_reduced.py cd ../../.. python example/image-classification/train_cifar10.py --network vgg16_reduced ` This PR will fix the above crash. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > mxnet mkl-dnn crashes for Max pooling with convention "full" > - > > Key: MXNET-33 > URL: https://issues.apache.org/jira/browse/MXNET-33 > Project: Apache MXNet > Issue Type: Bug >Reporter: Ashok Emani >Assignee: Ashok Emani >Priority: Minor > > A crash is seen when running SSD-VGG16 300 VOC , with mxnet mkldnn build, for > the following op. > > "op": "Pooling", > "attrs": { > "kernel": "(2, 2)", > "pool_type": "max", > "pooling_convention": "full", > "stride": "(2, 2)" > }, > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Commented] (MXNET-33) mxnet mkl-dnn crashes for Max pooling with convention "full"
[ https://issues.apache.org/jira/browse/MXNET-33?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16389103#comment-16389103 ] ASF GitHub Bot commented on MXNET-33: - piiswrong commented on a change in pull request #10021: [MXNET-33] Mkldnn pooling convention crash URL: https://github.com/apache/incubator-mxnet/pull/10021#discussion_r172750399 ## File path: src/operator/nn/mkldnn/mkldnn_pooling-inl.h ## @@ -92,6 +92,8 @@ inline bool SupportMKLDNNPooling(const PoolingParam , if (param.pooling_convention == pool_enum::kValid) return true; + else +return false; Review comment: this makes the later code unreachable. Are you sure this is what you want? This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > mxnet mkl-dnn crashes for Max pooling with convention "full" > - > > Key: MXNET-33 > URL: https://issues.apache.org/jira/browse/MXNET-33 > Project: Apache MXNet > Issue Type: Bug >Reporter: Ashok Emani >Assignee: Chris Olivier >Priority: Minor > > A crash is seen when running SSD-VGG16 300 VOC , with mxnet mkldnn build, for > the following op. > > "op": "Pooling", > "attrs": { > "kernel": "(2, 2)", > "pool_type": "max", > "pooling_convention": "full", > "stride": "(2, 2)" > }, > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Commented] (MXNET-33) mxnet mkl-dnn crashes for Max pooling with convention "full"
[ https://issues.apache.org/jira/browse/MXNET-33?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16389094#comment-16389094 ] ASF GitHub Bot commented on MXNET-33: - TaoLv commented on issue #10021: [MXNET-33] Mkldnn pooling convention crash URL: https://github.com/apache/incubator-mxnet/pull/10021#issuecomment-371034685 There is a test case for max pooling with `full` convention: https://github.com/apache/incubator-mxnet/blob/master/tests/python/gpu/test_operator_gpu.py#L841 But it isn't observed that this case failed before. If the test case doesn't cover the corner case you mentioned, maybe we need add new test case for it. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > mxnet mkl-dnn crashes for Max pooling with convention "full" > - > > Key: MXNET-33 > URL: https://issues.apache.org/jira/browse/MXNET-33 > Project: Apache MXNet > Issue Type: Bug >Reporter: Ashok Emani >Assignee: Chris Olivier >Priority: Minor > > A crash is seen when running SSD-VGG16 300 VOC , with mxnet mkldnn build, for > the following op. > > "op": "Pooling", > "attrs": { > "kernel": "(2, 2)", > "pool_type": "max", > "pooling_convention": "full", > "stride": "(2, 2)" > }, > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Commented] (MXNET-33) mxnet mkl-dnn crashes for Max pooling with convention "full"
[ https://issues.apache.org/jira/browse/MXNET-33?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16389065#comment-16389065 ] ASF GitHub Bot commented on MXNET-33: - ashokei commented on issue #10021: [MXNET-33] Mkldnn pooling convention crash URL: https://github.com/apache/incubator-mxnet/pull/10021#issuecomment-371031468 @zheng-da @TaoLv please review. I notice we do not have proper "full" pooling convention support, this PR disables mkl-dnn pooling for "full" convention; We will probably need a new PR to implement the "full" convention feature with mkl-dnn for both Max and Avg Pooling cases. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > mxnet mkl-dnn crashes for Max pooling with convention "full" > - > > Key: MXNET-33 > URL: https://issues.apache.org/jira/browse/MXNET-33 > Project: Apache MXNet > Issue Type: Bug >Reporter: Ashok Emani >Assignee: Chris Olivier >Priority: Minor > > A crash is seen when running SSD-VGG16 300 VOC , with mxnet mkldnn build, for > the following op. > > "op": "Pooling", > "attrs": { > "kernel": "(2, 2)", > "pool_type": "max", > "pooling_convention": "full", > "stride": "(2, 2)" > }, > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org
[jira] [Commented] (MXNET-33) mxnet mkl-dnn crashes for Max pooling with convention "full"
[ https://issues.apache.org/jira/browse/MXNET-33?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16389048#comment-16389048 ] ASF GitHub Bot commented on MXNET-33: - ashokei opened a new pull request #10021: [MXNET-33] Mkldnn pooling convention crash URL: https://github.com/apache/incubator-mxnet/pull/10021 ## Description ## As crash is seen if pooling convention is "full" with Max pool. This PR resolves the crash by defaulting to CPU implementation when pooling convention is "full". ## Checklist ## ### Essentials ### - [ ✅ ] Passed code style checking (`make lint`) - [ ✅ ] Changes are complete (i.e. I finished coding on this PR) ### Changes ### - disable mkl-dnn if pooling convention is "full" ## Comments ## - This PR is to resolve the crash seen. A new PR may be needed to implement "full" pooling convention with mkl-dnn. This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > mxnet mkl-dnn crashes for Max pooling with convention "full" > - > > Key: MXNET-33 > URL: https://issues.apache.org/jira/browse/MXNET-33 > Project: Apache MXNet > Issue Type: Bug >Reporter: Ashok Emani >Assignee: Chris Olivier >Priority: Minor > > A crash is seen when running SSD-VGG16 300 VOC , with mxnet mkldnn build, for > the following op. > > "op": "Pooling", > "attrs": { > "kernel": "(2, 2)", > "pool_type": "max", > "pooling_convention": "full", > "stride": "(2, 2)" > }, > -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: issues-unsubscr...@mxnet.apache.org For additional commands, e-mail: issues-h...@mxnet.apache.org