You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@singa.apache.org by GitBox <gi...@apache.org> on 2020/08/30 04:12:35 UTC
[GitHub] [singa] nudles opened a new issue #784: Preparation for V3.1 release
nudles opened a new issue #784:
URL: https://github.com/apache/singa/issues/784
I propose to release a minor version to reflect the changes since v3.0, including
Please test the following items and check the documentation if they are done.
- [ ] new onnx models and operators @joddiy
- [ ] distributed training @chrishkchris
- [ ] computational graph to support RNN @chrishkchris
- [ ] website update @nudles
- [ ] pypi package generation @nudles
- [ ] github workflow for code quality and coverage management @moazreyad
- [ ] tensor APIs @dcslin
- [ ] new autograd operators @joddiy
- [ ] anything else?
Here is the checklist and [steps](http://www.apache.org/dev/release-publishing.html)
- [ ] Select a release manager. The release manager (RM) is the coordinator for the release process. It is the RM's signature (.asc) that is uploaded together with the release. The RM generates KEY (RSA 4096-bit) and uploads it to a public key server. The RM needs to get his key endorsed (signed) by other Apache user, to be connected to the web of trust. He should first ask the mentor to help signing his key. [How to generate the key](http://www.apache.org/dev/release-signing.html)?
- [ ] Check license. [FAQ](https://www.apache.org/legal/src-headers.html#faq-docs); [SINGA Issue](https://issues.apache.org/jira/projects/SINGA/issues/SINGA-447)
- [ ] the codebase does not include third-party code which is not compatible to APL;
- [ ] The dependencies are compatible with APL. GNU-like licenses are NOT compatible;
- [ ] All source files written by us MUST include the Apache license header: http://www.apache.org/legal/src-headers.html. There's a script in there which helps propagating the header to all files.
- [ ] Update the LICENSE file. If we include any third party code in the release package which is not APL, must state it at the end of the [LICENSE](https://github.com/apache/singa/blob/master/LICENSE#L448) file and include the license boilerplate in the original file.
- [ ] Bump the version. Check code and documentation
- [ ] The build process is error-free.
- [ ] Unit tests are included (as much as possible)
- [ ] Conda packages run without errors.
- [ ] The online documentation on the Apache website is up to date.
- [ ] Prepare the RELEASE_NOTES file. Include the following items, Introduction, Features, Bugs (link to JIRA or Github PR), Changes, Dependency list, Incompatibility issues. Follow this [example](http://commons.apache.org/proper/commons-digester/commons-digester-3.0/RELEASE-NOTES.txt).
- [ ] Prepare DISCLAIMER file. Modify from the [template](http://incubator.apache.org/guides/branding.html#disclaimers)
- [ ] Package the release candidate. The release should be packaged into : apache-singa-VERSION.tar.gz. The release should not include any binary files including git files. Upload the release to for [stage](https://dist.apache.org/repos/dist/dev/VERSION/). The tar file, signature, KEY and SHA256 checksum file should be included. MD5 is no longer used. Policy is [here](http://www.apache.org/dev/release-distribution#sigs-and-sums)
* apache-singa-VERSION.tar.gz
* KEY
* XX.acs
* .SHA256
- [ ] Call for vote by sending an email
```
To: dev@singa.apache.org
Subject: [VOTE] Release apache-singa-X.Y.Z (release candidate N)
Hi all,
I have created a build for Apache SINGA X.Y.Z, release candidate N.
The artifacts to be voted on are located here: xxxx
The hashes of the artifacts are as follows: xxx
Release artifacts are signed with the following key: xxx
Please vote on releasing this package. The vote is open for at least 72 hours and passes if a majority of at least three +1 votes are cast.
[ ] +1 Release this package as Apache SINGA X.Y.Z
[ ] 0 I don't feel strongly about it, but I'm okay with the release
[ ] -1 Do not release this package because...
Here is my vote:
+1
```
- [ ] Wait at least 48 hours for test responses. Any PMC, committer or contributor can test features for releasing, and feedback. Everyone should check these before vote +1. If the vote passes, then send the result email. Otherwise, repeat from the beginning.
```
Subject: [RESULT] [VOTE] Release apache-singa-X.Y.Z (release candidate N)
To: dev@singa.apache.org
Thanks to everyone who has voted and given their comments. The tally is as follows.
N binding +1s:
<names>
N non-binding +1s:
<names>
No 0s or -1s.
I am delighted to announce that the proposal to release Apache SINGA X.Y.Zhas passed.
````
- [ ] Upload the package for [distribution](http://www.apache.org/dev/release-publishing.html#distribution) to https://dist.apache.org/repos/dist/release/VERSION/.
- [ ] Update the Download page of SINGA website. The tar.gz file MUST be downloaded from mirror, using closer.cgi script; other artifacts MUST be downloaded from main Apache site. More details [here](http://www.apache.org/dev/release-download-pages.html). Some feedback we got during the previous releases: "Download pages must only link to formal releases, so must not include links to GitHub.", "Links to KEYS, sigs and hashes must not use dist.apache.org; instead use https://www.apache.org/dist/singa/...;", "Also you only need one KEYS link, and there should be a description of how to use KEYS + sig or hash to verify the downloads."
- [ ] Remove the RC tag and compile the conda packages.
- [ ] Publish the release information.
```
To: announce@apache.org, dev@singa.apache.org
Subject: [ANNOUNCE] Apache SINGA X.Y.Z released
We are pleased to announce that SINGA X.Y.Z is released.
SINGA is a general distributed deep learning platform for training big deep learning models over large datasets.
The release is available at: http://singa.apache.org/downloads.html
The main features of this release include XXX
We look forward to hearing your feedback, suggestions, and contributions to the project.
On behalf of the SINGA team, {SINGA Team Member Name}
````
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] moazreyad commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
moazreyad commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-731728636
Great. I updated the infobox of the wikipedia page. I think we can close the issue now.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris closed issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris closed issue #784:
URL: https://github.com/apache/singa/issues/784
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] moazreyad commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
moazreyad commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-687988720
The workflow for coverage is ready in #788 and Travis CI can be removed now as discussed in #790.
More workflows can be made to automate packaging (pypi, debian, etc.) and to automate website updates.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-683378516
got it, thanks! I will test the distributed training and computational graph to support RNN, also check the corresponding documentation.
update on 9/1: submitted some fix to https://github.com/apache/singa/pull/785 , some test result of the distributed training is pasted on that PR
update on 9/3: worked with Rulin @XJDKC to fix the graph operation in https://github.com/apache/singa/pull/787
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-683378516
got it, thanks! I will test the distributed training and computational graph to support RNN, also check the corresponding documentation.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-683378516
got it, thanks! I will test the distributed training and computational graph to support RNN, also check the corresponding documentation.
update on 9/1: submitted some fix to https://github.com/apache/singa/pull/785 , some test result of the distributed training is pasted on that PR
updated on 9/3: worked with Rulin @XJDKC to fix the graph operation in https://github.com/apache/singa/pull/787
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-683378516
got it, thanks! I will test the distributed training and computational graph to support RNN, also check the corresponding documentation.
update on 9/1: submitted some fix to https://github.com/apache/singa/pull/785 , some test result of the distributed training is pasted on that PR
update on 9/3: worked with Rulin @XJDKC to fix the graph operation in https://github.com/apache/singa/pull/787
update on 9/8: updated the documentation of dist train and comp graph on PR https://github.com/apache/singa-doc/pull/27
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] dcslin commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
dcslin commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698093681
changelog 2020-09-24:
- Added
- BiLSTM model on InsuranceQA example training script.
- Tensor transformation (reshape, transpose) supports up to 6 dimensions.
- Implemented traverse_unary_transform in Cuda backend, which is similar to CPP backend one.
- Added tensor operation erf, rounde (round even).
- Added some sanity check on autograd input to prevent fatal error caused by unexpected input shape.
- Changed
- Fix IMDB LSTM model example training script.
- Fix Tensor operation Mult on Broadcasting use cases.
- Gaussian function on Tensor now can run on Tensor with odd size.
- Updated a testing helper function gradients() in autograd to lookup param gradient by param python object id for testing purpose.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] joddiy removed a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
joddiy removed a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698963047
- Add new autograd and ONNX operators: Erf, Embedding, Rounde, Round, Where, SpaceToDepth, DepthToSpace, UpSample, Pad, Expand, CosSim, Floor, Ceil.
- Add new layers: Embedding, Gemm.
- Reconstruct `sonnx`:
- Support creating operators from both layer and autograd.
- Re-write `SingaRep` to provide more powerful intermediate representation of SINGA.
- Add a `SONNXModel` which implements from 'Model' to provide uniform API and features.
- Add new ONNX models: SqueezeNet, ShuffleNetv1, ShuffleNetv2, VGG19, DenseNet121, GPT2, RoBERTa.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-683378516
got it, thanks! I will test the distributed training and computational graph to support RNN, also check the corresponding documentation.
update on 9/1: submitted some fix to https://github.com/apache/singa/pull/785 , some test result of the distributed training is pasted on that 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:
users@infra.apache.org
[GitHub] [singa] moazreyad closed issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
moazreyad closed issue #784:
URL: https://github.com/apache/singa/issues/784
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] dcslin edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
dcslin edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698093681
* BiLSTM model on [InsuranceQA](https://github.com/shuzi/insuranceQA) example training script. The baseline model is inspired by Tan, Ming, et al. "Lstm-based deep learning models for non-factoid answer selection." arXiv preprint arXiv:1511.04108 (2015).
* Tensor Refactoring and Enhancement
- Tensor transformation (reshape, transpose) supports up to 6 dimensions.
- Implemented traverse_unary_transform in Cuda backend, which is similar to CPP backend one.
- Added tensor operation erf, rounde (round even).
- Fix Tensor operation Mult on Broadcasting use cases.
- Gaussian function on Tensor now can run on Tensor with odd size.
* autograd
- Added some sanity check on autograd input to prevent fatal error caused by unexpected input shape.
- Updated a testing helper function gradients() in autograd to lookup param gradient by param python object id for testing purpose.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] nudles commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
nudles commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698088998
Please add details about the changes that you have made since v3.0.0. Follow the [release note](https://github.com/apache/singa/blob/master/RELEASE_NOTES)
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] joddiy edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
joddiy edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698968614
* 12 new operators are added into the autograd module: CosSim, DepthToSpace, Embedding, Erf, Expand, Floor, Pad, Round, Rounde, SpaceToDepth, UpSample, Where.
* 2 layers are added into layer module: Embedding, Gemm.
* 9 new operators are added to sonnx module for both backend and frontend:
[DepthToSpace](https://github.com/onnx/onnx/blob/master/docs/Operators.md#DepthToSpace),
[Erf](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Erf),
[Expand](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Expand),
[Floor](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Floor),
[Pad](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Pad),
[Round](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Round),
[SpaceToDepth](https://github.com/onnx/onnx/blob/master/docs/Operators.md#SpaceToDepth),
[UpSample](https://github.com/onnx/onnx/blob/master/docs/Operators.md#UpSample),
[Where](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Where),
Their tests are added as well.
* Some ONNX models are imported into SINGA, including
[DenseNet121](https://github.com/onnx/models/blob/master/vision/classification/densenet-121),
[ShuffleNetv1](https://github.com/onnx/models/blob/master/vision/classification/shufflenet),
[ShuffleNetv2](https://github.com/onnx/models/blob/master/vision/classification/shufflenet),
[SqueezeNet](https://github.com/onnx/models/blob/master/vision/classification/squeezenet),
[VGG19](https://github.com/onnx/models/blob/master/vision/classification/vgg),
[GPT2](https://github.com/onnx/models/blob/master/text/machine_comprehension/gpt-2),
[RoBERTa](https://github.com/onnx/models/blob/master/text/machine_comprehension/roberta),
* Reconstruct soonx,
- Support creating operators from both `layer` and `autograd`.
- Re-write `SingaRep` to provide a more powerful intermediate representation of SINGA.
- Add a `SONNXModel` which implements from `Model` to provide uniform API and features.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] joddiy commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
joddiy commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698968614
* 12 new operators are added into the autograd module: Erf, Embedding, Rounde, Round, Where, SpaceToDepth, DepthToSpace, UpSample, Pad, Expand, CosSim, Floor.
* 2 layers are added into layer module: Embedding, Gemm.
* 10 new operators are added to sonnx module for both backend and frontend:
[Erf](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Erf),
[Embedding](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Embedding),
[Round](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Round),
[Where](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Where),
[SpaceToDepth](https://github.com/onnx/onnx/blob/master/docs/Operators.md#SpaceToDepth),
[DepthToSpace](https://github.com/onnx/onnx/blob/master/docs/Operators.md#DepthToSpace),
[UpSample](https://github.com/onnx/onnx/blob/master/docs/Operators.md#UpSample),
[Pad](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Pad),
[Expand](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Expand),
[Floor](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Floor),
Their tests are added as well.
* Some ONNX models are imported into SINGA, including
[SqueezeNet](https://github.com/onnx/models/blob/master/vision/classification/squeezenet),
[ShuffleNetv1](https://github.com/onnx/models/blob/master/vision/classification/shufflenet),
[ShuffleNetv2](https://github.com/onnx/models/blob/master/vision/classification/shufflenet),
[VGG19](https://github.com/onnx/models/blob/master/vision/classification/vgg),
[DenseNet121](https://github.com/onnx/models/blob/master/vision/classification/densenet-121),
[GPT2](https://github.com/onnx/models/blob/master/text/machine_comprehension/gpt-2),
[RoBERTa](https://github.com/onnx/models/blob/master/text/machine_comprehension/roberta),
* Reconstruct soonx,
- Support creating operators from both layer and autograd.
- Re-write `SingaRep` to provide more powerful intermediate representation of SINGA.
- Add a `SONNXModel` which implements from 'Model' to provide uniform API and features.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] nudles commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
nudles commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-731705319
I have sent the announcement email.
Thanks for the reminder.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-727160609
3.1 already released
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] moazreyad commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
moazreyad commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-731634597
> 3.1 already released
I see that there is [vote result email in singa-dev mail list](https://mail-archives.apache.org/mod_mbox/singa-dev/202010.mbox/%3CCAJz0iLunmJ5pqwggCejG2OgzehvLF%3DogEqDFC-kPziV%2BPB%3D37w%40mail.gmail.com%3E). However, I don't see the `Publish the release information.` step.
An email "[ANNOUNCE] Apache SINGA X.Y.Z released" should be sent to announce@apache.org, dev@singa.apache.org as given above in the check list of release tasks. This step is required to complete the release process.
Note also that the email body says "The release is available at: http://singa.apache.org/downloads.html", however this page does not exist in the new website. We need to have downloads.html page because it is like a standard in Apache projects to have ProjectName.apache.org/downloads.html . The page `http://singa.apache.org/docs/download-singa/` does not follow the standard URLs of the Apache projects, and it also does not contain the 3.1 release.
Finally, it will be nice to update the [wikipedia page of singa](https://en.wikipedia.org/wiki/Apache_SINGA) with the new version information.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] nudles commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
nudles commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698088998
Please add details about the changes that you have made since v3.0.0. Follow the [release note](https://github.com/apache/singa/blob/master/RELEASE_NOTES)
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] joddiy commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
joddiy commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698963047
- Add new autograd and ONNX operators: Erf, Embedding, Rounde, Round, Where, SpaceToDepth, DepthToSpace, UpSample, Pad, Expand, CosSim, Floor, Ceil.
- Add new layers: Embedding, Gemm.
- Reconstruct `sonnx`:
- Support creating operators from both layer and autograd.
- Re-write `SingaRep` to provide more powerful intermediate representation of SINGA.
- Add a `SONNXModel` which implements from 'Model' to provide uniform API and features.
- Add new ONNX models: SqueezeNet, ShuffleNetv1, ShuffleNetv2, VGG19, DenseNet121, GPT2, RoBERTa.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-683378516
got it, thanks! I will test the distributed training and computational graph to support RNN, also check the corresponding documentation.
update on 9/1: submitted some fix to https://github.com/apache/singa/pull/785 , some test result of the distributed training is pasted on that PR
update on 9/3: worked with Rulin @XJDKC to fix the graph operation in https://github.com/apache/singa/pull/787
update on 9/8: updated the documentation of dist train and comp graph on PR https://github.com/apache/singa-doc/pull/27
update on 9/27: added the documentation of Optimizer, Time Profiling function, and Model Checkpoint Save Load Function on PR https://github.com/apache/singa-doc/pull/31
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] joddiy edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
joddiy edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698968614
* 12 new operators are added into the autograd module: CosSim, DepthToSpace, Embedding, Erf, Expand, Floor, Pad, Round, Rounde, SpaceToDepth, UpSample, Where.
* 2 layers are added into layer module: Embedding, Gemm.
* 9 new operators are added to sonnx module for both backend and frontend:
[DepthToSpace](https://github.com/onnx/onnx/blob/master/docs/Operators.md#DepthToSpace),
[Erf](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Erf),
[Expand](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Expand),
[Floor](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Floor),
[Pad](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Pad),
[Round](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Round),
[SpaceToDepth](https://github.com/onnx/onnx/blob/master/docs/Operators.md#SpaceToDepth),
[UpSample](https://github.com/onnx/onnx/blob/master/docs/Operators.md#UpSample),
[Where](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Where),
Their tests are added as well.
* Some ONNX models are imported into SINGA, including
[DenseNet121](https://github.com/onnx/models/blob/master/vision/classification/densenet-121),
[ShuffleNetv1](https://github.com/onnx/models/blob/master/vision/classification/shufflenet),
[ShuffleNetv2](https://github.com/onnx/models/blob/master/vision/classification/shufflenet),
[SqueezeNet](https://github.com/onnx/models/blob/master/vision/classification/squeezenet),
[VGG19](https://github.com/onnx/models/blob/master/vision/classification/vgg),
[GPT2](https://github.com/onnx/models/blob/master/text/machine_comprehension/gpt-2),
[RoBERTa](https://github.com/onnx/models/blob/master/text/machine_comprehension/roberta),
* Reconstruct soonx,
- Support creating operators from both layer and autograd.
- Re-write `SingaRep` to provide more powerful intermediate representation of SINGA.
- Add a `SONNXModel` which implements from 'Model' to provide uniform API and features.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] dcslin commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
dcslin commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698093681
changelog 2020-09-24:
- Added
- BiLSTM model on InsuranceQA example training script.
- Tensor transformation (reshape, transpose) supports up to 6 dimensions.
- Implemented traverse_unary_transform in Cuda backend, which is similar to CPP backend one.
- Added tensor operation erf, rounde (round even).
- Added some sanity check on autograd input to prevent fatal error caused by unexpected input shape.
- Changed
- Fix IMDB LSTM model example training script.
- Fix Tensor operation Mult on Broadcasting use cases.
- Gaussian function on Tensor now can run on Tensor with odd size.
- Updated a testing helper function gradients() in autograd to lookup param gradient by param python object id for testing purpose.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] dcslin edited a comment on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
dcslin edited a comment on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-698093681
* BiLSTM model on [InsuranceQA](https://github.com/shuzi/insuranceQA) example training script. The baseline model is inspired by Tan, Ming, et al. "Lstm-based deep learning models for non-factoid answer selection." arXiv preprint arXiv:1511.04108 (2015).
* Tensor Refactoring and Enhancement
- Tensor transformation (reshape, transpose) supports up to 6 dimensions.
- Implemented traverse_unary_transform in Cuda backend, which is similar to CPP backend one.
- Added tensor operation erf, rounde (round even).
- Fix Tensor operation Mult on Broadcasting use cases.
- Gaussian function on Tensor now can run on Tensor with odd size.
* autograd
- Added some sanity check on autograd input to prevent fatal error caused by unexpected input shape.
- Updated a testing helper function gradients() in autograd to lookup param gradient by param python object id for testing purpose.
----------------------------------------------------------------
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:
users@infra.apache.org
[GitHub] [singa] chrishkchris commented on issue #784: Preparation for V3.1 release
Posted by GitBox <gi...@apache.org>.
chrishkchris commented on issue #784:
URL: https://github.com/apache/singa/issues/784#issuecomment-731699784
Now I added a redirect link so that http://singa.apache.org/downloads.html is working now. It also has the 3.1 release. Does it look better now?
> Note also that the email body says "The release is available at: http://singa.apache.org/downloads.html", however this page does not exist in the new website. We need to have downloads.html page because it is like a standard in Apache projects to have ProjectName.apache.org/downloads.html . The page http://singa.apache.org/docs/download-singa/ does not follow the standard URLs of the Apache projects, and it also does not contain the 3.1 release.
----------------------------------------------------------------
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:
users@infra.apache.org