Skip to content
This repository was archived by the owner on Nov 17, 2023. It is now read-only.

[NumPy] turn on set_np #18631

Closed
szha wants to merge 2 commits into from
Closed

[NumPy] turn on set_np #18631

szha wants to merge 2 commits into from

Conversation

@szha
Copy link
Member

@szha szha commented Jun 28, 2020

Description

(Brief description on what this PR is about)

Checklist

Essentials

Please feel free to remove inapplicable items for your PR.

  • The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant JIRA issue created (except PRs with tiny changes)
  • Changes are complete (i.e. I finished coding on this PR)
  • All changes have test coverage:
  • Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
  • Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
  • Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
  • Code is well-documented:
  • For user-facing API changes, API doc string has been updated.
  • For new C++ functions in header files, their functionalities and arguments are documented.
  • For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
  • Check the API doc at https://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
  • To the best of my knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change

Changes

  • Feature1, tests, (and when applicable, API doc)
  • Feature2, tests, (and when applicable, API doc)

Comments

  • If this change is a backward incompatible change, why must this change be made.
  • Interesting edge cases to note here

@mxnet-bot
Copy link

mxnet-bot commented Jun 28, 2020

Hey @szha , Thanks for submitting the PR
All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the following commands:

  • To trigger all jobs: @mxnet-bot run ci [all]
  • To trigger specific jobs: @mxnet-bot run ci [job1, job2]

CI supported jobs: [sanity, clang, windows-gpu, website, centos-gpu, miscellaneous, edge, unix-cpu, windows-cpu, unix-gpu, centos-cpu]

Note:
Only following 3 categories can trigger CI :PR Author, MXNet Committer, Jenkins Admin.
All CI tests must pass before the PR can be merged.

@leezu leezu mentioned this pull request Jul 13, 2020
Merged
4 tasks
@szha szha mentioned this pull request Apr 20, 2021
Open
6 tasks
@barry-jin barry-jin mentioned this pull request May 20, 2021
Merged
49 tasks
@szha szha closed this May 21, 2021
@szha szha deleted the np_default branch May 21, 2021 20:10
leezu pushed a commit that referenced this pull request Jun 21, 2021
@barry-jin
Switch all HybridBlocks to use forward interface (#20262)
7152685 
Changes

Adopt packed_func based ffi on npx.group_norm
Gluon2.0 upgrade: use forward interface in blocks

    gluon/data/vision/*
    gluon/loss.py
    gluon/model_zoo/*
    gluon/nn/*
    gluon/rnn/*

use np/npx interface in gluon/metric.py
Implement infer_shape method

    gluon/nn/basic_layers.py::Dense
    gluon/nn/basic_layers.py::BatchNorm
    gluon/nn/basic_layers.py::InstanceNorm
    gluon/nn/basic_layers.py::LayerNorm
    gluon/nn/basic_layers.py::GroupNorm
    gluon/nn/conv_layers.py::_Conv
    gluon/nn/conv_layers.py::DeformableConvolution
    gluon/nn/conv_layers.py::ModulatedDeformableConvolution

Remove hybrid mode with F in gluon/probability/*
gluon/rnn/*

    conv_rnn_cell.py: implement forward, infer_shape, np/npx
    rnn_cell.py: implement forward, np/npx; use special infer_shape method based on layer, input_size and if it's bidirectional.
    rnn_layer.py: implement forward, infer_shape, np/npx

fix issue np.average return ADT type; npx.pooling
Copy control flow ops(loop_while, cond, foreach) from ndarray.contrib to mx.npx.control_flow
Register some legacy ops in npx

    stes_op
    sync_batch_norm
    legacy pad (np.pad doesn't have backward computation for 'reflect" mode)
    Some rnn related: sequence_last, sequence_reverse, slice_channel, broadcast_greater, softsign

Use forward, np/npx for all the gluon related tests; remove gluon tests with symbol inputs

    remove test_gluon_data_vision.py and test_gluon_probability_v1.py as related tests are covered in test_numpy_gluon_data_vision.py and test_gluon_probability_v2.py
    Test test_numpy_op.py::test_np_nan_to_num only for copy argument is set to True, since Inplace operations are not supported when recording in deferred compute mode

Remove hybrid_block interface in gluon/block.py
Remove hybrid_block interface in documentation and docstring

    update docs/python_docs/python/tutorials/packages/gluon/blocks/custom_layers
    remove python_tutorials/packages/gluon/blocks/custom_layers_beginners.md as it's duplicate to custom_layers
    remove docs/python_docs/python/docstutorials/packages/legacy/ndarray/sparse/train_gluon as gluon2.0 do not support sparse
    update docs/python_docs/python/tutorials/packages/gluon/blocks/custom_loss
    update docs/python_docs/python/tutorials/packages/gluon/blocks/hybridize

Turn on NumPy mode by default (#18631)
Fix gluon2.0 reference leak.
Migrate control flow operators to npx namespace

    Foreach
    while_loop
    cond
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@szha @mxnet-bot