[GitHub] [incubator-mxnet] szha commented on issue #18662: out of memory issue while using mxnet with sockeye
szha commented on issue #18662: URL: https://github.com/apache/incubator-mxnet/issues/18662#issuecomment-654032873 @MrRaghav thanks for creating the issue. What model of GPU are you using? What's the GPU memory size? Also, have you tried using `export MXNET_GPU_MEM_POOL_TYPE=Round`? https://mxnet.apache.org/api/faq/env_var#memory-options ``` Round: A memory pool that always rounds the requested memory size and allocates memory of the rounded size. MXNET_GPU_MEM_POOL_ROUND_LINEAR_CUTOFF defines how to round up a memory size. Caching and allocating buffered memory works in the same way as the naive memory pool. ``` This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] szha edited a comment on issue #18662: out of memory issue while using mxnet with sockeye
szha edited a comment on issue #18662: URL: https://github.com/apache/incubator-mxnet/issues/18662#issuecomment-654032873 @MrRaghav thanks for creating the issue. What model of GPU are you using? What's the GPU memory size? Also, have you tried using `export MXNET_GPU_MEM_POOL_TYPE=Round`? https://mxnet.apache.org/api/faq/env_var#memory-options Round: A memory pool that always rounds the requested memory size and allocates memory of the rounded size. MXNET_GPU_MEM_POOL_ROUND_LINEAR_CUTOFF defines how to round up a memory size. Caching and allocating buffered memory works in the same way as the naive memory pool. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] bgawrych commented on pull request #18158: [numpy] fix npx.softmax for 0-sized inputs
bgawrych commented on pull request #18158: URL: https://github.com/apache/incubator-mxnet/pull/18158#issuecomment-654098082 @Yiyan66 @yzhliu Do you think this fix should be cherry-picked to 1.x? Without it my change https://github.com/apache/incubator-mxnet/pull/18602 is failing with floating point exception This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[incubator-mxnet] tag 1.7.0.rc0 created (now 477affe)
This is an automated email from the ASF dual-hosted git repository. taolv pushed a change to tag 1.7.0.rc0 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. at 477affe (commit) No new revisions were added by this update.
[incubator-mxnet] tag 1.7.0.rc0 created (now 477affe)
This is an automated email from the ASF dual-hosted git repository. taolv pushed a change to tag 1.7.0.rc0 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. at 477affe (commit) No new revisions were added by this update.
[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.
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 3eee77c Bump the publish timestamp. 3eee77c is described below commit 3eee77cb368c653d0f8302bd7e00621648832e29 Author: mxnet-ci AuthorDate: Mon Jul 6 06:42:25 2020 + Bump the publish timestamp. --- date.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/date.txt b/date.txt new file mode 100644 index 000..b905f13 --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Mon Jul 6 06:42:25 UTC 2020
[incubator-mxnet-site] branch asf-site updated: Publish triggered by CI
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 a2004b4 Publish triggered by CI a2004b4 is described below commit a2004b4e20339789e72d340dd3b4f89de4d9709a Author: mxnet-ci AuthorDate: Mon Jul 6 06:42:20 2020 + Publish triggered by CI --- date.txt | 1 - feed.xml | 2 +- 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/date.txt b/date.txt deleted file mode 100644 index 947d918..000 --- a/date.txt +++ /dev/null @@ -1 +0,0 @@ -Mon Jul 6 00:41:17 UTC 2020 diff --git a/feed.xml b/feed.xml index 5f92c71..649ef1f 100644 --- a/feed.xml +++ b/feed.xml @@ -1 +1 @@ -http://www.w3.org/2005/Atom; >https://jekyllrb.com/; version="4.0.0">Jekyllhttps://mxnet.apache.org/feed.xml; rel="self" type="application/atom+xml" />https://mxnet.apache.org/; rel="alternate" type="text/html" />2020-07-06T00:30:31+00:00https://mxnet.apache.org/feed.xmlApache MXNetA flexible and efficient library for deep [...] \ No newline at end of file +http://www.w3.org/2005/Atom; >https://jekyllrb.com/; version="4.0.0">Jekyllhttps://mxnet.apache.org/feed.xml; rel="self" type="application/atom+xml" />https://mxnet.apache.org/; rel="alternate" type="text/html" />2020-07-06T06:31:09+00:00https://mxnet.apache.org/feed.xmlApache MXNetA flexible and efficient library for deep [...] \ No newline at end of file
[GitHub] [incubator-mxnet] cenggokhan closed issue #18030: Model load problem in 64 bit devices
cenggokhan closed issue #18030: URL: https://github.com/apache/incubator-mxnet/issues/18030 This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] cosmincatalin commented on issue #18655: MXNet for Scala 2.12 and 2.13
cosmincatalin commented on issue #18655: URL: https://github.com/apache/incubator-mxnet/issues/18655#issuecomment-654082313 I've tinkered a little with building from source, but wasn't very successful, I guess I need to focus on it more. To answer your second question, yes, I am interested specifically in the CPU packages. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[incubator-mxnet] tag 1.7.0.rc0 created (now 477affe)
This is an automated email from the ASF dual-hosted git repository. taolv pushed a change to tag 1.7.0.rc0 in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git. at 477affe (commit) No new revisions were added by this update.
[GitHub] [incubator-mxnet] MrRaghav commented on issue #18662: out of memory issue while using mxnet with sockeye
MrRaghav commented on issue #18662: URL: https://github.com/apache/incubator-mxnet/issues/18662#issuecomment-654201954 Hello, please find the information in following points - 1) I am using **rtx2080ti** 2) To run sockeye, I used 3 GPUs and specified device ids. Memory of GPUs is as follows _username@server:~/username/sockeye$ nvidia-smi --format=csv --query-gpu=memory.total memory.total [MiB] 11019 MiB 11019 MiB 11019 MiB_ 3) regarding export command, I tried running sockeye.train like below: _export MXNET_GPU_MEM_POOL_TYPE=Round_ _python3 -m sockeye.train -s trained.BPE.de -t trained.BPE.en -vs dev.BPE.de -vt dev.BPE.en --shared-vocab \ --device-ids -3 --max-checkpoints 3 -o model_ But, still got the same error. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[incubator-mxnet-site] branch asf-site updated: Publish triggered by CI
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 c4ac656 Publish triggered by CI c4ac656 is described below commit c4ac65623517f92420946b682f0ae2601a76792f Author: mxnet-ci AuthorDate: Mon Jul 6 12:53:31 2020 + Publish triggered by CI --- date.txt | 1 - feed.xml | 2 +- 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/date.txt b/date.txt deleted file mode 100644 index b905f13..000 --- a/date.txt +++ /dev/null @@ -1 +0,0 @@ -Mon Jul 6 06:42:25 UTC 2020 diff --git a/feed.xml b/feed.xml index 649ef1f..b09155a 100644 --- a/feed.xml +++ b/feed.xml @@ -1 +1 @@ -http://www.w3.org/2005/Atom; >https://jekyllrb.com/; version="4.0.0">Jekyllhttps://mxnet.apache.org/feed.xml; rel="self" type="application/atom+xml" />https://mxnet.apache.org/; rel="alternate" type="text/html" />2020-07-06T06:31:09+00:00https://mxnet.apache.org/feed.xmlApache MXNetA flexible and efficient library for deep [...] \ No newline at end of file +http://www.w3.org/2005/Atom; >https://jekyllrb.com/; version="4.0.0">Jekyllhttps://mxnet.apache.org/feed.xml; rel="self" type="application/atom+xml" />https://mxnet.apache.org/; rel="alternate" type="text/html" />2020-07-06T12:30:16+00:00https://mxnet.apache.org/feed.xmlApache MXNetA flexible and efficient library for deep [...] \ No newline at end of file
[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.
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 382e599 Bump the publish timestamp. 382e599 is described below commit 382e59901c02ba3e6a54d3561ad549b89dbe9dad Author: mxnet-ci AuthorDate: Mon Jul 6 12:53:36 2020 + Bump the publish timestamp. --- date.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/date.txt b/date.txt new file mode 100644 index 000..9c29e97 --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Mon Jul 6 12:53:36 UTC 2020
svn commit: r40320 - in /dev/incubator/mxnet/1.7.0.rc0: ./ apache-mxnet-src-1.7.0.rc0-incubating.tar.gz apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.asc apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.sh
Author: taolv Date: Mon Jul 6 13:26:45 2020 New Revision: 40320 Log: Add mxnet-1.7.0.rc0 Added: dev/incubator/mxnet/1.7.0.rc0/ dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz (with props) dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.asc (with props) dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.sha512 Added: dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz == Binary file - no diff available. Propchange: dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz -- svn:mime-type = application/x-gzip Added: dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.asc == Binary file - no diff available. Propchange: dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.asc -- svn:mime-type = application/pgp-signature Added: dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.sha512 == --- dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.sha512 (added) +++ dev/incubator/mxnet/1.7.0.rc0/apache-mxnet-src-1.7.0.rc0-incubating.tar.gz.sha512 Mon Jul 6 13:26:45 2020 @@ -0,0 +1 @@ +67401ff5d0ed3e84cbf82e7d2903caea07c21627c1dcf7af708a74b8e85ecb1c36df37bf20c2aa1666bbe71dcae855b5c3e7863de6abccd3a89516c5e3ea58da apache-mxnet-src-1.7.0.rc0-incubating.tar.gz
[GitHub] [incubator-mxnet] Yiyan66 commented on pull request #18660: [numpy][do not review] fix flaky mixed precision binary error
Yiyan66 commented on pull request #18660: URL: https://github.com/apache/incubator-mxnet/pull/18660#issuecomment-654217844 @mxnet-bot run ci [centos-cpu] This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18660: [numpy][do not review] fix flaky mixed precision binary error
mxnet-bot commented on pull request #18660: URL: https://github.com/apache/incubator-mxnet/pull/18660#issuecomment-654217914 Jenkins CI successfully triggered : [centos-cpu] This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] Kh4L commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
Kh4L commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450375033 ## File path: python/mxnet/gluon/block.py ## @@ -1040,41 +1040,62 @@ def _build_cache(self, *args): warnings.warn("Parameter %s is not used by any computation. " "Is this intended?"%unused, stacklevel=4) -data_indices = [] -param_indices = [] -self._cached_op_args = [] -for i, name in enumerate(input_names): -if name in data_names: -data_indices.append(i) -self._cached_op_args.append((True, data_names[name])) -else: -param_indices.append(i) -self._cached_op_args.append((False, params[name])) -flags = [('data_indices', data_indices), ('param_indices', param_indices)] + \ -self._flags - args, _ = _flatten(args, "input") try: -for is_arg, i in self._cached_op_args: -if not is_arg: -i.data() +for name in input_names: +if name in params: +params[name].data() except DeferredInitializationError: self._deferred_infer_shape(*args) -for is_arg, i in self._cached_op_args: -if not is_arg: -i._finish_deferred_init() +for name in input_names: +if name in params: +params[name]._finish_deferred_init() +arg_dict, aux_dict = dict(), dict() if self._backend: ctx = args[0].context # get list of params in the order of out.list_arguments -arg_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_arguments()} -aux_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_auxiliary_states()} +arg_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_arguments()}) +aux_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_auxiliary_states()}) # Partition the graph. out = out.optimize_for(self._backend, arg_dict, aux_dict, ctx, **self._backend_opts) + #update cached graph with partitioned graph self._cached_graph = data, out + +input_names = out.list_inputs() +data_indices = [] +param_indices = [] +self._cached_op_args = [] +for i, name in enumerate(input_names): +pair = None +if name in data_names: +data_indices.append(i) +pair = (True, data_names[name]) +else: +param_indices.append(i) +if name in params: +param = params[name] +else: +assert self._backend, "Parameter " + name + " is missing from block params" +if name in arg_dict or name: +param_data = arg_dict[name] +elif name in aux_dict: +param_data = aux_dict[name] +else: +raise RuntimeError('Expected inputs missing from arg and aux after partioning. ' Review comment: Done This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] szha commented on issue #18662: out of memory issue while using mxnet with sockeye
szha commented on issue #18662: URL: https://github.com/apache/incubator-mxnet/issues/18662#issuecomment-654383639 @fhieber do you have recommendation on how to run sockeye on the above GPU? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] leezu commented on a change in pull request #18625: Enable Large Tensor Support: Stage 1
leezu commented on a change in pull request #18625: URL: https://github.com/apache/incubator-mxnet/pull/18625#discussion_r450407053 ## File path: ci/docker/runtime_functions.sh ## @@ -512,6 +519,7 @@ build_ubuntu_gpu_clang10_werror() { -DMXNET_CUDA_ARCH="$CI_CMAKE_CUDA_ARCH" \ -DCMAKE_BUILD_TYPE="RelWithDebInfo" \ -DUSE_CPP_PACKAGE=OFF \ + -DUSE_INT64_TENSOR_SIZE=OFF \ Review comment: What's the reason for splitting this up into multiple stages? The current PR is only changing the default settings (enable the setting by default). As long as there are bugs in various environments (eg with clang), you will break users. Thus I recommend including all required fixes in this PR. Also note that this clang build here has nothing to do with `MKL_USE_ILP64` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-mxnet] leezu commented on a change in pull request #18625: Enable Large Tensor Support: Stage 1
leezu commented on a change in pull request #18625: URL: https://github.com/apache/incubator-mxnet/pull/18625#discussion_r450407053 ## File path: ci/docker/runtime_functions.sh ## @@ -512,6 +519,7 @@ build_ubuntu_gpu_clang10_werror() { -DMXNET_CUDA_ARCH="$CI_CMAKE_CUDA_ARCH" \ -DCMAKE_BUILD_TYPE="RelWithDebInfo" \ -DUSE_CPP_PACKAGE=OFF \ + -DUSE_INT64_TENSOR_SIZE=OFF \ Review comment: What's the reason for splitting this up into multiple stages? The current PR is only changing the default settings (enable the setting by default). As long as there are bugs in various environments (eg with clang), you will break users. Thus I recommend including all required fixes in this PR. Also note that this clang build here is unrelated to `MKL_USE_ILP64`. Thanks for your work on the large tensor support! This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] leezu commented on a change in pull request #18625: Enable Large Tensor Support: Stage 1
leezu commented on a change in pull request #18625: URL: https://github.com/apache/incubator-mxnet/pull/18625#discussion_r450407053 ## File path: ci/docker/runtime_functions.sh ## @@ -512,6 +519,7 @@ build_ubuntu_gpu_clang10_werror() { -DMXNET_CUDA_ARCH="$CI_CMAKE_CUDA_ARCH" \ -DCMAKE_BUILD_TYPE="RelWithDebInfo" \ -DUSE_CPP_PACKAGE=OFF \ + -DUSE_INT64_TENSOR_SIZE=OFF \ Review comment: What's the reason for splitting this up into multiple stages? The current PR is only changing the default settings (enable the setting by default). As long as there are bugs in various environments (eg with clang), you will break users, because these users won't know that the default changed and that they now need to opt into the small tensor case. Thus I recommend including all required fixes in this PR. Also note that this clang build here is unrelated to `MKL_USE_ILP64`. Thanks for your work on the large tensor support! This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.
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 641764b Bump the publish timestamp. 641764b is described below commit 641764b0e6e990b0ec872aaf7755a23036b521cb Author: mxnet-ci AuthorDate: Mon Jul 6 18:40:53 2020 + Bump the publish timestamp. --- date.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/date.txt b/date.txt new file mode 100644 index 000..8f55ea6 --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Mon Jul 6 18:40:53 UTC 2020
[incubator-mxnet-site] branch asf-site updated: Publish triggered by CI
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 d2bacf1 Publish triggered by CI d2bacf1 is described below commit d2bacf109db931e22da112204965542f378c29e5 Author: mxnet-ci AuthorDate: Mon Jul 6 18:40:48 2020 + Publish triggered by CI --- date.txt | 1 - feed.xml | 2 +- 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/date.txt b/date.txt deleted file mode 100644 index 9c29e97..000 --- a/date.txt +++ /dev/null @@ -1 +0,0 @@ -Mon Jul 6 12:53:36 UTC 2020 diff --git a/feed.xml b/feed.xml index b09155a..4a19274 100644 --- a/feed.xml +++ b/feed.xml @@ -1 +1 @@ -http://www.w3.org/2005/Atom; >https://jekyllrb.com/; version="4.0.0">Jekyllhttps://mxnet.apache.org/feed.xml; rel="self" type="application/atom+xml" />https://mxnet.apache.org/; rel="alternate" type="text/html" />2020-07-06T12:30:16+00:00https://mxnet.apache.org/feed.xmlApache MXNetA flexible and efficient library for deep [...] \ No newline at end of file +http://www.w3.org/2005/Atom; >https://jekyllrb.com/; version="4.0.0">Jekyllhttps://mxnet.apache.org/feed.xml; rel="self" type="application/atom+xml" />https://mxnet.apache.org/; rel="alternate" type="text/html" />2020-07-06T18:30:18+00:00https://mxnet.apache.org/feed.xmlApache MXNetA flexible and efficient library for deep [...] \ No newline at end of file
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450346162 ## File path: python/mxnet/gluon/block.py ## @@ -1040,41 +1040,62 @@ def _build_cache(self, *args): warnings.warn("Parameter %s is not used by any computation. " "Is this intended?"%unused, stacklevel=4) -data_indices = [] -param_indices = [] -self._cached_op_args = [] -for i, name in enumerate(input_names): -if name in data_names: -data_indices.append(i) -self._cached_op_args.append((True, data_names[name])) -else: -param_indices.append(i) -self._cached_op_args.append((False, params[name])) -flags = [('data_indices', data_indices), ('param_indices', param_indices)] + \ -self._flags - args, _ = _flatten(args, "input") try: -for is_arg, i in self._cached_op_args: -if not is_arg: -i.data() +for name in input_names: +if name in params: +params[name].data() except DeferredInitializationError: self._deferred_infer_shape(*args) -for is_arg, i in self._cached_op_args: -if not is_arg: -i._finish_deferred_init() +for name in input_names: +if name in params: +params[name]._finish_deferred_init() +arg_dict, aux_dict = dict(), dict() if self._backend: ctx = args[0].context # get list of params in the order of out.list_arguments -arg_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_arguments()} -aux_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_auxiliary_states()} +arg_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_arguments()}) +aux_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_auxiliary_states()}) # Partition the graph. out = out.optimize_for(self._backend, arg_dict, aux_dict, ctx, **self._backend_opts) + #update cached graph with partitioned graph self._cached_graph = data, out + +input_names = out.list_inputs() +data_indices = [] +param_indices = [] +self._cached_op_args = [] +for i, name in enumerate(input_names): +pair = None +if name in data_names: +data_indices.append(i) +pair = (True, data_names[name]) +else: +param_indices.append(i) Review comment: nit: can we move this above the if/else since we do it in both? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] rondogency commented on issue #18665: Using ShiftScale operation (Similar to BatchNorm) in MXNet (Customer question/inquiry)
rondogency commented on issue #18665: URL: https://github.com/apache/incubator-mxnet/issues/18665#issuecomment-654368718 Hi, are you using Gluon API? according to the doc it is supported https://github.com/apache/incubator-mxnet/blob/v1.7.x/python/mxnet/gluon/nn/basic_layers.py#L306, if it's not then it would be a bug in Gluon Also which MXNet version are you using? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] Kh4L commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
Kh4L commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450376590 ## File path: python/mxnet/gluon/block.py ## @@ -1437,6 +1460,23 @@ def hybrid_forward(self, F, x, *args, **kwargs): # pylint: disable= invalid-name raise NotImplementedError +def reset_ctx(self, ctx): +"""Re-assign all Parameters to other contexts. If the Block is hybridized, it will reset the _cached_op_args. + +Parameters +-- +ctx : Context or list of Context, default :py:meth:`context.current_context()`. +Assign Parameter to given context. If ctx is a list of Context, a +copy will be made for each context. +""" +params = self.collect_params() +if self._cached_op: +for p in self._cached_op_args: +# resetting parameters creating by the partitioning backend +if p.name not in params: +p.reset_ctx(ctx) +for p in params.values(): Review comment: > Although i guess if we delete a param, then it will still be in params but not in _cached_op_args. That is the reason why We don't want to do any additional reset_ctx: reset is costly because it copies NDArrays I will add some comments to clarify This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] sammieghabra commented on issue #18665: Using ShiftScale operation (Similar to BatchNorm) in MXNet (Customer question/inquiry)
sammieghabra commented on issue #18665: URL: https://github.com/apache/incubator-mxnet/issues/18665#issuecomment-654371413 @rondogency Thank you so much for your response. As mentioned in the description of this issue, I am aware of the `use_global_stats` flag in batch norm. However, as I stated in the description, running mean and running var do not get updated if that flag is true for [training](https://github.com/apache/incubator-mxnet/blob/master/src/operator/nn/batch_norm.cc#L260-L271). For our case, we need those values to be updated for training. I am using version 1.5 of MXNet This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] ptrendx commented on pull request #18622: [WIP] Use RTC for elementwise and broadcast ops
ptrendx commented on pull request #18622: URL: https://github.com/apache/incubator-mxnet/pull/18622#issuecomment-654373793 @leezu Thanks for listing the files - I am now looking into the contents of the right cross compilation package. I think you are right that everything should already be there and it is possibly just a package that was not installed in the Dockerfile itself, so I should be able to do this in the PR :-). This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] stu1130 commented on pull request #18504: [Improvement] Invoke mkldnn and cudnn BatchNorm when axis != 1
stu1130 commented on pull request #18504: URL: https://github.com/apache/incubator-mxnet/pull/18504#issuecomment-654391355 @TaoLv could you help merge it if the PR looks good to you? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #18504: [Improvement] Invoke mkldnn and cudnn BatchNorm when axis != 1
mxnet-bot commented on pull request #18504: URL: https://github.com/apache/incubator-mxnet/pull/18504#issuecomment-654390942 None of the jobs entered are supported. Jobs entered by user: [macosx-x86_64] CI supported Jobs: [centos-gpu, miscellaneous, windows-cpu, windows-gpu, website, clang, unix-gpu, centos-cpu, sanity, unix-cpu, edge] This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] stu1130 commented on pull request #18504: [Improvement] Invoke mkldnn and cudnn BatchNorm when axis != 1
stu1130 commented on pull request #18504: URL: https://github.com/apache/incubator-mxnet/pull/18504#issuecomment-654390908 @mxnet-bot run ci [macosx-x86_64] This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450358036 ## File path: python/mxnet/gluon/block.py ## @@ -1437,6 +1460,23 @@ def hybrid_forward(self, F, x, *args, **kwargs): # pylint: disable= invalid-name raise NotImplementedError +def reset_ctx(self, ctx): +"""Re-assign all Parameters to other contexts. If the Block is hybridized, it will reset the _cached_op_args. + +Parameters +-- +ctx : Context or list of Context, default :py:meth:`context.current_context()`. +Assign Parameter to given context. If ctx is a list of Context, a +copy will be made for each context. +""" +params = self.collect_params() +if self._cached_op: +for p in self._cached_op_args: +# resetting parameters creating by the partitioning backend +if p.name not in params: +p.reset_ctx(ctx) +for p in params.values(): Review comment: Although i guess if we delete a param, then it will still be in `params` but not in `_cached_op_args`. And the context check will fail if we dont do all the params, so i guess this makes sense. Maybe we should have a comment that `params` and `_cached_op_args` might contain unique params (ie. there is no superset)? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450345405 ## File path: python/mxnet/gluon/block.py ## @@ -1040,41 +1040,62 @@ def _build_cache(self, *args): warnings.warn("Parameter %s is not used by any computation. " "Is this intended?"%unused, stacklevel=4) -data_indices = [] -param_indices = [] -self._cached_op_args = [] -for i, name in enumerate(input_names): -if name in data_names: -data_indices.append(i) -self._cached_op_args.append((True, data_names[name])) -else: -param_indices.append(i) -self._cached_op_args.append((False, params[name])) -flags = [('data_indices', data_indices), ('param_indices', param_indices)] + \ -self._flags - args, _ = _flatten(args, "input") try: -for is_arg, i in self._cached_op_args: -if not is_arg: -i.data() +for name in input_names: +if name in params: +params[name].data() except DeferredInitializationError: self._deferred_infer_shape(*args) -for is_arg, i in self._cached_op_args: -if not is_arg: -i._finish_deferred_init() +for name in input_names: +if name in params: +params[name]._finish_deferred_init() +arg_dict, aux_dict = dict(), dict() if self._backend: ctx = args[0].context # get list of params in the order of out.list_arguments -arg_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_arguments()} -aux_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_auxiliary_states()} +arg_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_arguments()}) Review comment: makes sense, thanks! This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450348025 ## File path: python/mxnet/gluon/block.py ## @@ -1040,41 +1040,62 @@ def _build_cache(self, *args): warnings.warn("Parameter %s is not used by any computation. " "Is this intended?"%unused, stacklevel=4) -data_indices = [] -param_indices = [] -self._cached_op_args = [] -for i, name in enumerate(input_names): -if name in data_names: -data_indices.append(i) -self._cached_op_args.append((True, data_names[name])) -else: -param_indices.append(i) -self._cached_op_args.append((False, params[name])) -flags = [('data_indices', data_indices), ('param_indices', param_indices)] + \ -self._flags - args, _ = _flatten(args, "input") try: -for is_arg, i in self._cached_op_args: -if not is_arg: -i.data() +for name in input_names: +if name in params: +params[name].data() except DeferredInitializationError: self._deferred_infer_shape(*args) -for is_arg, i in self._cached_op_args: -if not is_arg: -i._finish_deferred_init() +for name in input_names: +if name in params: +params[name]._finish_deferred_init() +arg_dict, aux_dict = dict(), dict() if self._backend: ctx = args[0].context # get list of params in the order of out.list_arguments -arg_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_arguments()} -aux_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_auxiliary_states()} +arg_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_arguments()}) +aux_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_auxiliary_states()}) # Partition the graph. out = out.optimize_for(self._backend, arg_dict, aux_dict, ctx, **self._backend_opts) + #update cached graph with partitioned graph self._cached_graph = data, out + +input_names = out.list_inputs() +data_indices = [] +param_indices = [] +self._cached_op_args = [] +for i, name in enumerate(input_names): +pair = None +if name in data_names: +data_indices.append(i) +pair = (True, data_names[name]) +else: +param_indices.append(i) +if name in params: +param = params[name] +else: +assert self._backend, "Parameter " + name + " is missing from block params" Review comment: Is this the case that should never happen? When a param name is not in `params` but `_backend` is not set? If so can we change the error message and maybe leave a comment? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450352450 ## File path: python/mxnet/gluon/block.py ## @@ -1437,6 +1460,23 @@ def hybrid_forward(self, F, x, *args, **kwargs): # pylint: disable= invalid-name raise NotImplementedError +def reset_ctx(self, ctx): +"""Re-assign all Parameters to other contexts. If the Block is hybridized, it will reset the _cached_op_args. + +Parameters +-- +ctx : Context or list of Context, default :py:meth:`context.current_context()`. +Assign Parameter to given context. If ctx is a list of Context, a +copy will be made for each context. +""" +params = self.collect_params() +if self._cached_op: +for p in self._cached_op_args: +# resetting parameters creating by the partitioning backend +if p.name not in params: +p.reset_ctx(ctx) +for p in params.values(): Review comment: Can you explain why we need to loop over `_cached_op_args` and only reset the params not in `params`, and then loop again over `params` and then reset them there instead? Is it possible to do the work in the 2nd loop in the first loop? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] chinakook commented on issue #18646: BatchNorm with axis=-1 is much slower than axis=1
chinakook commented on issue #18646: URL: https://github.com/apache/incubator-mxnet/issues/18646#issuecomment-654296913 I think NHWC layout is very important in point cloud algorithms. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450349091 ## File path: python/mxnet/gluon/block.py ## @@ -1040,41 +1040,62 @@ def _build_cache(self, *args): warnings.warn("Parameter %s is not used by any computation. " "Is this intended?"%unused, stacklevel=4) -data_indices = [] -param_indices = [] -self._cached_op_args = [] -for i, name in enumerate(input_names): -if name in data_names: -data_indices.append(i) -self._cached_op_args.append((True, data_names[name])) -else: -param_indices.append(i) -self._cached_op_args.append((False, params[name])) -flags = [('data_indices', data_indices), ('param_indices', param_indices)] + \ -self._flags - args, _ = _flatten(args, "input") try: -for is_arg, i in self._cached_op_args: -if not is_arg: -i.data() +for name in input_names: +if name in params: +params[name].data() except DeferredInitializationError: self._deferred_infer_shape(*args) -for is_arg, i in self._cached_op_args: -if not is_arg: -i._finish_deferred_init() +for name in input_names: +if name in params: +params[name]._finish_deferred_init() +arg_dict, aux_dict = dict(), dict() if self._backend: ctx = args[0].context # get list of params in the order of out.list_arguments -arg_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_arguments()} -aux_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_auxiliary_states()} +arg_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_arguments()}) +aux_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_auxiliary_states()}) # Partition the graph. out = out.optimize_for(self._backend, arg_dict, aux_dict, ctx, **self._backend_opts) + #update cached graph with partitioned graph self._cached_graph = data, out + +input_names = out.list_inputs() +data_indices = [] +param_indices = [] +self._cached_op_args = [] +for i, name in enumerate(input_names): +pair = None +if name in data_names: +data_indices.append(i) +pair = (True, data_names[name]) +else: +param_indices.append(i) +if name in params: +param = params[name] +else: +assert self._backend, "Parameter " + name + " is missing from block params" +if name in arg_dict or name: +param_data = arg_dict[name] +elif name in aux_dict: +param_data = aux_dict[name] +else: +raise RuntimeError('Expected inputs missing from arg and aux after partioning. ' Review comment: Is this the case when the backend added a param to the graph but not to the arg/aux dict? If so, can we change the error message to say something like "param was added to the graph but the tensor was not added to args/aux"? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] leezu commented on pull request #18622: [WIP] Use RTC for elementwise and broadcast ops
leezu commented on pull request #18622: URL: https://github.com/apache/incubator-mxnet/pull/18622#issuecomment-654353049 @ptrendx these are the files available in the S3 bucket. You may need to edit the Dockerfile to install one of the debs: ``` % aws s3 ls --recursive s3://mxnet-ci-prod-private-slave-data/nvidia 6s ~ 2020-04-11 02:46:49 0 nvidia/ 2020-04-11 02:47:05 0 nvidia/sdkm_downloads/ 2020-04-11 03:07:04 237414931 nvidia/sdkm_downloads/Jetson-210_Linux_R32.3.1_aarch64.tbz2 2020-04-11 03:03:33 288481398 nvidia/sdkm_downloads/NVIDIA_Nsight_Graphics_L4T_2019.5.19322.deb 2020-04-11 03:24:005350765 nvidia/sdkm_downloads/NVIDIA_VisionWorks_References.zip 2020-04-11 03:17:26 113581742 nvidia/sdkm_downloads/NsightSystems-linux-public-2019.6.2.6-3ffb807.deb 2020-04-11 03:24:081022926 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-dev.deb 2020-04-11 03:23:599920774 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-libs.deb 2020-04-11 03:24:13 14240 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-licenses.deb 2020-04-11 03:24:072350538 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-python.deb 2020-04-11 03:24:11 192038 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-samples.deb 2020-04-11 03:24:091027754 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-dev.deb 2020-04-11 03:21:58 45701978 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-libs.deb 2020-04-11 03:24:12 18306 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-licenses.deb 2020-04-11 03:24:032576742 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-python.deb 2020-04-11 03:24:11 192030 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-samples.deb 2020-04-11 03:24:11 143245 nvidia/sdkm_downloads/SDKM_logs_JetPack_4.3_Linux_for_Jetson_TX1_2020-04-10_17-43-23.zip 2020-04-11 02:51:29 1263228889 nvidia/sdkm_downloads/Tegra_Linux_Sample-Root-Filesystem_R32.3.1_aarch64.tbz2 2020-04-11 02:51:29 394475774 nvidia/sdkm_downloads/cuda-repo-cross-aarch64-10-0-local-10.0.326_1.0-1_all.deb 2020-04-11 02:51:29 1000333292 nvidia/sdkm_downloads/cuda-repo-l4t-10-0-local-10.0.326_1.0-1_arm64.deb 2020-04-11 02:51:29 1582667320 nvidia/sdkm_downloads/cuda-repo-ubuntu1804-10-0-local-10.0.326-410.108_1.0-1_amd64.deb 2020-04-11 03:22:38 25538902 nvidia/sdkm_downloads/devtools_docs.zip 2020-04-11 03:24:13 16108 nvidia/sdkm_downloads/graphsurgeon-tf_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:13:51 161324708 nvidia/sdkm_downloads/libcudnn7-dev_7.6.3.28-1+cuda10.0_arm64.deb 2020-04-11 03:23:556627420 nvidia/sdkm_downloads/libcudnn7-doc_7.6.3.28-1+cuda10.0_arm64.deb 2020-04-11 03:13:18 181721612 nvidia/sdkm_downloads/libcudnn7_7.6.3.28-1+cuda10.0_arm64.deb 2020-04-11 03:24:13 12596 nvidia/sdkm_downloads/libnvidia-container-tools_0.9.0_beta.1_arm64.deb 2020-04-11 03:24:12 42744 nvidia/sdkm_downloads/libnvidia-container0_0.9.0_beta.1_arm64.deb 2020-04-11 03:24:12 69900 nvidia/sdkm_downloads/libnvinfer-bin_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:20:36 57300292 nvidia/sdkm_downloads/libnvinfer-dev_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:024460940 nvidia/sdkm_downloads/libnvinfer-doc_6.0.1-1+cuda10.0_all.deb 2020-04-11 03:24:081546362 nvidia/sdkm_downloads/libnvinfer-plugin-dev_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:061489314 nvidia/sdkm_downloads/libnvinfer-plugin6_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 02:51:29 451737338 nvidia/sdkm_downloads/libnvinfer-samples_6.0.1-1+cuda10.0_all.deb 2020-04-11 03:21:10 56443400 nvidia/sdkm_downloads/libnvinfer6_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:11 164522 nvidia/sdkm_downloads/libnvonnxparsers-dev_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:09 578042 nvidia/sdkm_downloads/libnvonnxparsers6_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:10 538398 nvidia/sdkm_downloads/libnvparsers-dev_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:08 707550 nvidia/sdkm_downloads/libnvparsers6_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:19:50 66561472 nvidia/sdkm_downloads/libvisionworks-repo_1.6.0.500n_amd64.deb 2020-04-11 03:20:07 60780704 nvidia/sdkm_downloads/libvisionworks-repo_1.6.0.500n_arm64.deb 2020-04-11 03:22:59 20597586 nvidia/sdkm_downloads/libvisionworks-sfm-repo_0.90.4_amd64.deb 2020-04-11 03:23:08 20216850 nvidia/sdkm_downloads/libvisionworks-sfm-repo_0.90.4_arm64.deb 2020-04-11 03:23:29 13113238 nvidia/sdkm_downloads/libvisionworks-tracking-repo_0.88.2_amd64.deb 2020-04-11 03:23:56 13107966 nvidia/sdkm_downloads/libvisionworks-tracking-repo_0.88.2_arm64.deb 2020-04-11 03:24:13 14842
[GitHub] [incubator-mxnet] leezu edited a comment on pull request #18622: [WIP] Use RTC for elementwise and broadcast ops
leezu edited a comment on pull request #18622: URL: https://github.com/apache/incubator-mxnet/pull/18622#issuecomment-654353049 @ptrendx these are the files available in the S3 bucket. They are the files obtained from the SDK Manager. Is the file you are looking for included? If not, if you provide the file we can add it to the bucket. ``` % aws s3 ls --recursive s3://mxnet-ci-prod-private-slave-data/nvidia 6s ~ 2020-04-11 02:46:49 0 nvidia/ 2020-04-11 02:47:05 0 nvidia/sdkm_downloads/ 2020-04-11 03:07:04 237414931 nvidia/sdkm_downloads/Jetson-210_Linux_R32.3.1_aarch64.tbz2 2020-04-11 03:03:33 288481398 nvidia/sdkm_downloads/NVIDIA_Nsight_Graphics_L4T_2019.5.19322.deb 2020-04-11 03:24:005350765 nvidia/sdkm_downloads/NVIDIA_VisionWorks_References.zip 2020-04-11 03:17:26 113581742 nvidia/sdkm_downloads/NsightSystems-linux-public-2019.6.2.6-3ffb807.deb 2020-04-11 03:24:081022926 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-dev.deb 2020-04-11 03:23:599920774 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-libs.deb 2020-04-11 03:24:13 14240 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-licenses.deb 2020-04-11 03:24:072350538 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-python.deb 2020-04-11 03:24:11 192038 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-aarch64-samples.deb 2020-04-11 03:24:091027754 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-dev.deb 2020-04-11 03:21:58 45701978 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-libs.deb 2020-04-11 03:24:12 18306 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-licenses.deb 2020-04-11 03:24:032576742 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-python.deb 2020-04-11 03:24:11 192030 nvidia/sdkm_downloads/OpenCV-4.1.1-2-gd5a58aa75-x86_64-samples.deb 2020-04-11 03:24:11 143245 nvidia/sdkm_downloads/SDKM_logs_JetPack_4.3_Linux_for_Jetson_TX1_2020-04-10_17-43-23.zip 2020-04-11 02:51:29 1263228889 nvidia/sdkm_downloads/Tegra_Linux_Sample-Root-Filesystem_R32.3.1_aarch64.tbz2 2020-04-11 02:51:29 394475774 nvidia/sdkm_downloads/cuda-repo-cross-aarch64-10-0-local-10.0.326_1.0-1_all.deb 2020-04-11 02:51:29 1000333292 nvidia/sdkm_downloads/cuda-repo-l4t-10-0-local-10.0.326_1.0-1_arm64.deb 2020-04-11 02:51:29 1582667320 nvidia/sdkm_downloads/cuda-repo-ubuntu1804-10-0-local-10.0.326-410.108_1.0-1_amd64.deb 2020-04-11 03:22:38 25538902 nvidia/sdkm_downloads/devtools_docs.zip 2020-04-11 03:24:13 16108 nvidia/sdkm_downloads/graphsurgeon-tf_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:13:51 161324708 nvidia/sdkm_downloads/libcudnn7-dev_7.6.3.28-1+cuda10.0_arm64.deb 2020-04-11 03:23:556627420 nvidia/sdkm_downloads/libcudnn7-doc_7.6.3.28-1+cuda10.0_arm64.deb 2020-04-11 03:13:18 181721612 nvidia/sdkm_downloads/libcudnn7_7.6.3.28-1+cuda10.0_arm64.deb 2020-04-11 03:24:13 12596 nvidia/sdkm_downloads/libnvidia-container-tools_0.9.0_beta.1_arm64.deb 2020-04-11 03:24:12 42744 nvidia/sdkm_downloads/libnvidia-container0_0.9.0_beta.1_arm64.deb 2020-04-11 03:24:12 69900 nvidia/sdkm_downloads/libnvinfer-bin_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:20:36 57300292 nvidia/sdkm_downloads/libnvinfer-dev_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:024460940 nvidia/sdkm_downloads/libnvinfer-doc_6.0.1-1+cuda10.0_all.deb 2020-04-11 03:24:081546362 nvidia/sdkm_downloads/libnvinfer-plugin-dev_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:061489314 nvidia/sdkm_downloads/libnvinfer-plugin6_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 02:51:29 451737338 nvidia/sdkm_downloads/libnvinfer-samples_6.0.1-1+cuda10.0_all.deb 2020-04-11 03:21:10 56443400 nvidia/sdkm_downloads/libnvinfer6_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:11 164522 nvidia/sdkm_downloads/libnvonnxparsers-dev_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:09 578042 nvidia/sdkm_downloads/libnvonnxparsers6_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:10 538398 nvidia/sdkm_downloads/libnvparsers-dev_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:24:08 707550 nvidia/sdkm_downloads/libnvparsers6_6.0.1-1+cuda10.0_arm64.deb 2020-04-11 03:19:50 66561472 nvidia/sdkm_downloads/libvisionworks-repo_1.6.0.500n_amd64.deb 2020-04-11 03:20:07 60780704 nvidia/sdkm_downloads/libvisionworks-repo_1.6.0.500n_arm64.deb 2020-04-11 03:22:59 20597586 nvidia/sdkm_downloads/libvisionworks-sfm-repo_0.90.4_amd64.deb 2020-04-11 03:23:08 20216850 nvidia/sdkm_downloads/libvisionworks-sfm-repo_0.90.4_arm64.deb 2020-04-11 03:23:29 13113238 nvidia/sdkm_downloads/libvisionworks-tracking-repo_0.88.2_amd64.deb 2020-04-11 03:23:56 13107966
[GitHub] [incubator-mxnet] chinakook commented on issue #18591: CPU memory leak when running inference on model with GPU
chinakook commented on issue #18591: URL: https://github.com/apache/incubator-mxnet/issues/18591#issuecomment-654306719 This api may not support multi thread safety. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450346162 ## File path: python/mxnet/gluon/block.py ## @@ -1040,41 +1040,62 @@ def _build_cache(self, *args): warnings.warn("Parameter %s is not used by any computation. " "Is this intended?"%unused, stacklevel=4) -data_indices = [] -param_indices = [] -self._cached_op_args = [] -for i, name in enumerate(input_names): -if name in data_names: -data_indices.append(i) -self._cached_op_args.append((True, data_names[name])) -else: -param_indices.append(i) -self._cached_op_args.append((False, params[name])) -flags = [('data_indices', data_indices), ('param_indices', param_indices)] + \ -self._flags - args, _ = _flatten(args, "input") try: -for is_arg, i in self._cached_op_args: -if not is_arg: -i.data() +for name in input_names: +if name in params: +params[name].data() except DeferredInitializationError: self._deferred_infer_shape(*args) -for is_arg, i in self._cached_op_args: -if not is_arg: -i._finish_deferred_init() +for name in input_names: +if name in params: +params[name]._finish_deferred_init() +arg_dict, aux_dict = dict(), dict() if self._backend: ctx = args[0].context # get list of params in the order of out.list_arguments -arg_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_arguments()} -aux_dict = {name:args[data_names[name]] if name in data_names.keys() else params[name].data() -for name in out.list_auxiliary_states()} +arg_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_arguments()}) +aux_dict.update({name:args[data_names[name]] if name in data_names.keys() else params[name].data() + for name in out.list_auxiliary_states()}) # Partition the graph. out = out.optimize_for(self._backend, arg_dict, aux_dict, ctx, **self._backend_opts) + #update cached graph with partitioned graph self._cached_graph = data, out + +input_names = out.list_inputs() +data_indices = [] +param_indices = [] +self._cached_op_args = [] +for i, name in enumerate(input_names): +pair = None +if name in data_names: +data_indices.append(i) +pair = (True, data_names[name]) +else: +param_indices.append(i) Review comment: nit: can we move this above the if/else since we do it in both? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] xidulu commented on pull request #18403: Gluon.probability
xidulu commented on pull request #18403: URL: https://github.com/apache/incubator-mxnet/pull/18403#issuecomment-654344692 @leezu Issue solved by adding `garbage_expected` flag: https://github.com/apache/incubator-mxnet/pull/18403/files#diff-a6e158f7742ece308155a95c165505f6R2206 This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450360528 ## File path: python/mxnet/symbol/symbol.py ## @@ -1544,8 +1544,35 @@ def optimize_for(self, backend, args=None, aux=None, ctx=None, **kwargs): raise RuntimeError('Cannot add new aux in optimize_for since aux is None\n' + 'Provide a dictionary to the aux argument to optimize_for') -# return modified symbol -return Symbol(out) +new_sym = Symbol(out) + +arg_names = self.list_arguments() +new_arg_names = new_sym.list_arguments() +deleted_arg_names = set([item for item in arg_names + if item not in set(new_arg_names)]) + +if len(deleted_arg_names) > 0: +if args is not None: +for a_n in deleted_arg_names: +if a_n in args: +args.pop(a_n) +else: +warnings.warn('optimize_for deleted some argument. \n' + + 'Provide a dictionary to the arg argument to optimize_for') Review comment: Can we clarify this error message to something like: ``` A param was deleted during optimization, but no args dictionary was provided. Please ensure that your model weights match the newly optimized model. ``` Should we print the names of the deleted_arg_names in the message too (or is that overkill)? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450360528 ## File path: python/mxnet/symbol/symbol.py ## @@ -1544,8 +1544,35 @@ def optimize_for(self, backend, args=None, aux=None, ctx=None, **kwargs): raise RuntimeError('Cannot add new aux in optimize_for since aux is None\n' + 'Provide a dictionary to the aux argument to optimize_for') -# return modified symbol -return Symbol(out) +new_sym = Symbol(out) + +arg_names = self.list_arguments() +new_arg_names = new_sym.list_arguments() +deleted_arg_names = set([item for item in arg_names + if item not in set(new_arg_names)]) + +if len(deleted_arg_names) > 0: +if args is not None: +for a_n in deleted_arg_names: +if a_n in args: +args.pop(a_n) +else: +warnings.warn('optimize_for deleted some argument. \n' + + 'Provide a dictionary to the arg argument to optimize_for') Review comment: Can we clarify this error message to something like: ``` A param was deleted during optimization, but no args dictionary was provided. Please ensure that your model weights match the newly optimized model. ``` Should we print the names of the deleted_arg_names in the message too (or is that overkill)? This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] chinakook commented on issue #18643: ndarray.contrib.boolean_mask can not be hybridize
chinakook commented on issue #18643: URL: https://github.com/apache/incubator-mxnet/issues/18643#issuecomment-654293552 Actually, Loss is used explicitly, so it can be coded as Block, not HybridBlock. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] ptrendx commented on pull request #18622: [WIP] Use RTC for elementwise and broadcast ops
ptrendx commented on pull request #18622: URL: https://github.com/apache/incubator-mxnet/pull/18622#issuecomment-654316485 @ChaiBapchya The Jetson CI pipeline fails with not being able to find NVRTC - I confirmed it is part of jetpack, so I guess it is not included in the package of libs you have in S3. Can you update it? (And yes, I'm aware of issue #18637). This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] sammieghabra opened a new issue #18665: Using ShiftScale operation (Similar to BatchNorm) in MXNet (Customer question/inquiry)
sammieghabra opened a new issue #18665: URL: https://github.com/apache/incubator-mxnet/issues/18665 ## Description Hi MXNet team, My team wants to implement a layer in MXNet to implement a ShiftScale operator. This layer is similar to batchnorm - however, we want `moving_mean` and `moving_var` to be used instead of `data_mean` and `data_var` to compute the output of the layer. I see that batch norm has a flag `use_global_stats`, and in the [mxnet docs](http://beta.mxnet.io/r/api/mx.symbol.BatchNorm.html), it seems that setting this flag to be true would be something similar to what I'm trying to do. However, upon inspecting the [batch-norm code](https://github.com/apache/incubator-mxnet/blob/master/src/operator/nn/batch_norm.cc#L260-L271), it seems that running_mean and running_var won't be updated if that flag is set to true for training. 1. Is there a reason why from a design perspective setting this `use_global_stats` flag to be true wont update the running mean and running var? 2. We would like to support this shift scale layer during training. So what my proposal is to do is to add another flag to the `batchNorm` operator to be "use_shift_scale", which would simply replace mean and var with running mean and running var when updating the weights. Is this something that MXNet team would be ok with? 3. We also plan to train with more than one instance - will the running_mean and running_var parameters be the same across instances? Thanks Sammie This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] samskalicky commented on a change in pull request #18405: Add deleting of args aux aux to Partition API
samskalicky commented on a change in pull request #18405: URL: https://github.com/apache/incubator-mxnet/pull/18405#discussion_r450360818 ## File path: python/mxnet/symbol/symbol.py ## @@ -1544,8 +1544,35 @@ def optimize_for(self, backend, args=None, aux=None, ctx=None, **kwargs): raise RuntimeError('Cannot add new aux in optimize_for since aux is None\n' + 'Provide a dictionary to the aux argument to optimize_for') -# return modified symbol -return Symbol(out) +new_sym = Symbol(out) + +arg_names = self.list_arguments() +new_arg_names = new_sym.list_arguments() +deleted_arg_names = set([item for item in arg_names + if item not in set(new_arg_names)]) + +if len(deleted_arg_names) > 0: +if args is not None: +for a_n in deleted_arg_names: +if a_n in args: +args.pop(a_n) +else: +warnings.warn('optimize_for deleted some argument. \n' + + 'Provide a dictionary to the arg argument to optimize_for') +aux_names = self.list_auxiliary_states() +new_aux_names = new_sym.list_auxiliary_states() +deleted_aux_names = set([item for item in aux_names + if item not in set(new_aux_names)]) +if len(deleted_aux_names) > 0: +if aux is not None: +for a_n in deleted_aux_names: +if a_n in aux: +aux.pop(a_n) +else: +warnings.warn('optimize_for deleted some aux argument. \n' + + 'Provide a dictionary to the aux argument to optimize_for') Review comment: Same as above for args This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] leezu merged pull request #18639: User Feedback Widget
leezu merged pull request #18639: URL: https://github.com/apache/incubator-mxnet/pull/18639 This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] ys2843 commented on pull request #18639: User Feedback Widget
ys2843 commented on pull request #18639: URL: https://github.com/apache/incubator-mxnet/pull/18639#issuecomment-654522781 > @ys2843 can you please send an email to the mailing list requesting guidance on how to disclose the use of google analytics? @aaronmarkham made the point that the website is not compliant with GDPR. > We should ask ASF to consider including whatever guidance they give, into their policies. Will do. I will share the info with team later. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[incubator-mxnet] branch master updated: User Feedback Widget (#18639)
This is an automated email from the ASF dual-hosted git repository. lausen 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 54c0155 User Feedback Widget (#18639) 54c0155 is described below commit 54c0155b7581f5e10b1469a17ddf127d3c75e156 Author: Yang Shi AuthorDate: Mon Jul 6 17:01:42 2020 -0700 User Feedback Widget (#18639) * user feedback widget implementation * add user feedback widget to python docs site * update margin * add apache license * one more license * turn off feedback widget on python site * update copy * format * add event value field * turn on widget on Python site --- docs/python_docs/_static/feedback.css | 37 docs/python_docs/_static/feedback.js | 33 ++ .../themes/mx-theme/mxtheme/feedback.html | 10 ++ .../themes/mx-theme/mxtheme/layout.html| 5 ++- docs/static_site/src/_includes/feedback.html | 10 ++ docs/static_site/src/_includes/head.html | 3 ++ docs/static_site/src/_layouts/page_api.html| 3 ++ docs/static_site/src/_sass/feedback.scss | 39 ++ docs/static_site/src/assets/js/feedback.js | 33 ++ docs/static_site/src/assets/main.scss | 1 + 10 files changed, 173 insertions(+), 1 deletion(-) diff --git a/docs/python_docs/_static/feedback.css b/docs/python_docs/_static/feedback.css new file mode 100644 index 000..b4a64ec --- /dev/null +++ b/docs/python_docs/_static/feedback.css @@ -0,0 +1,37 @@ +.feedback-container { + text-align: center; +} + +.feedback-answer-container { + display: inline-block; +} + +.feedback-question { + display: inline-block; + padding: 0.5em 1em 0.5em 1em; +} + +.feedback-answer { + display: inline-block; + padding: 0.5em 1em 0.5em 1em; + color: #048ccc; + cursor: pointer; +} + +.feedback-answer:hover { + color: #ff; + background-color: #048ccc; +} + +.feedback-thank-you { + display: none; + padding: 0.5em 1em 0.5em 1em; +} + +.feedback-hr-top { + margin-top: 50px; +} + +.feedback-hr-bottom { + margin-bottom: 30px; +} diff --git a/docs/python_docs/_static/feedback.js b/docs/python_docs/_static/feedback.js new file mode 100644 index 000..f454237 --- /dev/null +++ b/docs/python_docs/_static/feedback.js @@ -0,0 +1,33 @@ +/*! + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ + +$(document).ready(function() { + $(".feedback-answer").on("click", function () { +$(".feedback-question").remove(); +$(".feedback-answer-container").remove(); +$(".feedback-thank-you").show(); +ga("send", { + hitType: "event", + eventCategory: "Did this page help you?", + eventAction: $(this).attr("data-response"), + eventLabel: window.location.pathname || "unknown", + eventValue: $(this).attr("data-response") === "yes" ? 1 : 0 +}); + }); +}); diff --git a/docs/python_docs/themes/mx-theme/mxtheme/feedback.html b/docs/python_docs/themes/mx-theme/mxtheme/feedback.html new file mode 100644 index 000..8a9b1b5 --- /dev/null +++ b/docs/python_docs/themes/mx-theme/mxtheme/feedback.html @@ -0,0 +1,10 @@ + + +Did this page help you? + +Yes +No + +Thanks for your feedback! + + diff --git a/docs/python_docs/themes/mx-theme/mxtheme/layout.html b/docs/python_docs/themes/mx-theme/mxtheme/layout.html index 03189c8..994da38 100644 --- a/docs/python_docs/themes/mx-theme/mxtheme/layout.html +++ b/docs/python_docs/themes/mx-theme/mxtheme/layout.html @@ -64,10 +64,12 @@ '_static/sphinx_materialdesign_theme.css', '_static/fontawesome/all.css', '_static/fonts.css', +'_static/feedback.css', ] %} {%- block header %} - + + {% endblock %} {%- block relbar1 %}{% endblock %} {%- block relbar2 %}{% include "relations.html" %}{% endblock %} @@ -85,6 +87,7 @@ {%- block document %} {% block body %} {% endblock %} +{% include "feedback.html" %}
[incubator-mxnet] branch master updated: User Feedback Widget (#18639)
This is an automated email from the ASF dual-hosted git repository. lausen 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 54c0155 User Feedback Widget (#18639) 54c0155 is described below commit 54c0155b7581f5e10b1469a17ddf127d3c75e156 Author: Yang Shi AuthorDate: Mon Jul 6 17:01:42 2020 -0700 User Feedback Widget (#18639) * user feedback widget implementation * add user feedback widget to python docs site * update margin * add apache license * one more license * turn off feedback widget on python site * update copy * format * add event value field * turn on widget on Python site --- docs/python_docs/_static/feedback.css | 37 docs/python_docs/_static/feedback.js | 33 ++ .../themes/mx-theme/mxtheme/feedback.html | 10 ++ .../themes/mx-theme/mxtheme/layout.html| 5 ++- docs/static_site/src/_includes/feedback.html | 10 ++ docs/static_site/src/_includes/head.html | 3 ++ docs/static_site/src/_layouts/page_api.html| 3 ++ docs/static_site/src/_sass/feedback.scss | 39 ++ docs/static_site/src/assets/js/feedback.js | 33 ++ docs/static_site/src/assets/main.scss | 1 + 10 files changed, 173 insertions(+), 1 deletion(-) diff --git a/docs/python_docs/_static/feedback.css b/docs/python_docs/_static/feedback.css new file mode 100644 index 000..b4a64ec --- /dev/null +++ b/docs/python_docs/_static/feedback.css @@ -0,0 +1,37 @@ +.feedback-container { + text-align: center; +} + +.feedback-answer-container { + display: inline-block; +} + +.feedback-question { + display: inline-block; + padding: 0.5em 1em 0.5em 1em; +} + +.feedback-answer { + display: inline-block; + padding: 0.5em 1em 0.5em 1em; + color: #048ccc; + cursor: pointer; +} + +.feedback-answer:hover { + color: #ff; + background-color: #048ccc; +} + +.feedback-thank-you { + display: none; + padding: 0.5em 1em 0.5em 1em; +} + +.feedback-hr-top { + margin-top: 50px; +} + +.feedback-hr-bottom { + margin-bottom: 30px; +} diff --git a/docs/python_docs/_static/feedback.js b/docs/python_docs/_static/feedback.js new file mode 100644 index 000..f454237 --- /dev/null +++ b/docs/python_docs/_static/feedback.js @@ -0,0 +1,33 @@ +/*! + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ + +$(document).ready(function() { + $(".feedback-answer").on("click", function () { +$(".feedback-question").remove(); +$(".feedback-answer-container").remove(); +$(".feedback-thank-you").show(); +ga("send", { + hitType: "event", + eventCategory: "Did this page help you?", + eventAction: $(this).attr("data-response"), + eventLabel: window.location.pathname || "unknown", + eventValue: $(this).attr("data-response") === "yes" ? 1 : 0 +}); + }); +}); diff --git a/docs/python_docs/themes/mx-theme/mxtheme/feedback.html b/docs/python_docs/themes/mx-theme/mxtheme/feedback.html new file mode 100644 index 000..8a9b1b5 --- /dev/null +++ b/docs/python_docs/themes/mx-theme/mxtheme/feedback.html @@ -0,0 +1,10 @@ + + +Did this page help you? + +Yes +No + +Thanks for your feedback! + + diff --git a/docs/python_docs/themes/mx-theme/mxtheme/layout.html b/docs/python_docs/themes/mx-theme/mxtheme/layout.html index 03189c8..994da38 100644 --- a/docs/python_docs/themes/mx-theme/mxtheme/layout.html +++ b/docs/python_docs/themes/mx-theme/mxtheme/layout.html @@ -64,10 +64,12 @@ '_static/sphinx_materialdesign_theme.css', '_static/fontawesome/all.css', '_static/fonts.css', +'_static/feedback.css', ] %} {%- block header %} - + + {% endblock %} {%- block relbar1 %}{% endblock %} {%- block relbar2 %}{% include "relations.html" %}{% endblock %} @@ -85,6 +87,7 @@ {%- block document %} {% block body %} {% endblock %} +{% include "feedback.html" %}
[GitHub] [incubator-mxnet] ys2843 closed issue #18548: Merge beta.mxnet.io To MXNet Official Website
ys2843 closed issue #18548: URL: https://github.com/apache/incubator-mxnet/issues/18548 This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] ys2843 commented on issue #18548: Merge beta.mxnet.io To MXNet Official Website
ys2843 commented on issue #18548: URL: https://github.com/apache/incubator-mxnet/issues/18548#issuecomment-654523856 Project completed. Closing the issue. 1. Created a new R docs website and merged to v1.6 website - Done 2. Created user feedback widget - Done 3. Redirect traffic from beta site to official website - Done This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[incubator-mxnet-site] branch asf-site updated: Bump the publish timestamp.
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 5b3f84a Bump the publish timestamp. 5b3f84a is described below commit 5b3f84a97a5983a4ed7d4d6fc127f9ee5d9e8fe6 Author: mxnet-ci AuthorDate: Tue Jul 7 00:50:32 2020 + Bump the publish timestamp. --- date.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/date.txt b/date.txt new file mode 100644 index 000..286c0ba --- /dev/null +++ b/date.txt @@ -0,0 +1 @@ +Tue Jul 7 00:50:32 UTC 2020
[GitHub] [incubator-mxnet] rondogency commented on issue #18665: Using ShiftScale operation (Similar to BatchNorm) in MXNet (Customer question/inquiry)
rondogency commented on issue #18665: URL: https://github.com/apache/incubator-mxnet/issues/18665#issuecomment-654546431 @zhreshold @sxjscience FYI This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] rondogency commented on issue #18665: Using ShiftScale operation (Similar to BatchNorm) in MXNet (Customer question/inquiry)
rondogency commented on issue #18665: URL: https://github.com/apache/incubator-mxnet/issues/18665#issuecomment-654546346 can confirm currently MXNet doesn't support both update global running mean/var and use running mean/var to calculate grads in one pass of Batchnorm, while Keras-MXNet can support it by passing mean/var as arguments to Batchnorm as a workaround adding the support would still have constraints, especially cudnn doesn't support it so it cannot run with the best performance; some other frameworks like caffe doesn't support this either it can be implemented by modifying batch_norm.cc/cu and add "use_shift_scale" option to control it This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] andrei5055 commented on pull request #18582: Refactoring of Pooled Storage Manager classes
andrei5055 commented on pull request #18582: URL: https://github.com/apache/incubator-mxnet/pull/18582#issuecomment-654562254 Here is first graph for GPU Storage Managers: ![image](https://user-images.githubusercontent.com/7293680/86691969-30b3c280-bfbe-11ea-9101-0e2ef18b2210.png) I will construct similar graphs for CPU, CPU_PINNED and Threaded Engines. My benchmarks are based on [that script](https://github.com/apache/incubator-mxnet/issues/17335). As you could see, some lines are almost identical, as it should be. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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
[GitHub] [incubator-mxnet] leezu opened a new issue #18666: Version selector install guides
leezu opened a new issue #18666: URL: https://github.com/apache/incubator-mxnet/issues/18666 ## Description Currently the get started page points to a single build from source guide. As the get started page will be shared among all releases (https://github.com/apache/incubator-mxnet/pull/18658), the selector should be updated to include the corresponding install guide based on the user's selection. ![image](https://user-images.githubusercontent.com/946903/86656357-444e3180-bf9c-11ea-933f-585db3660402.png) For example, if the user selects version 1.5.1 here, the recommended installation guide for 1.5.1 should be shown inline instead of referring to a general (all-versions) install guide. This is an automated message from the Apache Git Service. To respond to the message, please log on to 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