Giter VIP home page Giter VIP logo

Comments (10)

chrishkchris avatar chrishkchris commented on July 23, 2024

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 #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 #787

update on 9/8: updated the documentation of dist train and comp graph on PR apache/singa-doc#27

update on 9/27: added the documentation of Optimizer, Time Profiling function, and Model Checkpoint Save Load Function on PR apache/singa-doc#31

from singa.

moazreyad avatar moazreyad commented on July 23, 2024

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.

from singa.

nudles avatar nudles commented on July 23, 2024

Please add details about the changes that you have made since v3.0.0. Follow the release note

from singa.

dcslin avatar dcslin commented on July 23, 2024
  • BiLSTM model on 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.

from singa.

joddiy avatar joddiy commented on July 23, 2024
  • 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,
    Erf,
    Expand,
    Floor,
    Pad,
    Round,
    SpaceToDepth,
    UpSample,
    Where,
    Their tests are added as well.

  • Some ONNX models are imported into SINGA, including
    DenseNet121,
    ShuffleNetv1,
    ShuffleNetv2,
    SqueezeNet,
    VGG19,
    GPT2,
    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.

from singa.

chrishkchris avatar chrishkchris commented on July 23, 2024

3.1 already released

from singa.

moazreyad avatar moazreyad commented on July 23, 2024

3.1 already released

I see that there is vote result email in singa-dev mail list. However, I don't see the Publish the release information. step.

An email "[ANNOUNCE] Apache SINGA X.Y.Z released" should be sent to [email protected], [email protected] 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 with the new version information.

from singa.

chrishkchris avatar chrishkchris commented on July 23, 2024

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.

from singa.

nudles avatar nudles commented on July 23, 2024

I have sent the announcement email.
Thanks for the reminder.

from singa.

moazreyad avatar moazreyad commented on July 23, 2024

Great. I updated the infobox of the wikipedia page. I think we can close the issue now.

from singa.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.