Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Patch 1 #1

Merged
merged 411 commits into from
Jun 9, 2019
Merged

Patch 1 #1

merged 411 commits into from
Jun 9, 2019

Conversation

ymjiang
Copy link
Owner

@ymjiang ymjiang commented Jun 9, 2019

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 $JIRA_ID refers to the relevant JIRA issue created (except PRs with tiny changes)
  • Changes are complete (i.e. I finished coding on this PR)
  • All changes have test coverage:
  • Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
  • Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
  • Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
  • Code is well-documented:
  • For user-facing API changes, API doc string has been updated.
  • For new C++ functions in header files, their functionalities and arguments are documented.
  • For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
  • Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
  • To the my best 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

eric-haibin-lin and others added 30 commits November 14, 2018 15:06
* patch fix

* update ignore

* rename getContext to bindToDevice

* Update JavaBenchmark.java
* use cached version of get thread max

* reserve core affects omp singleton

* omp_thread_max_ updated in one line

* remove enabled block

* add brackets

* re-add excluded reserved

* add missing var

* refactor macro
* Addressing PR feedback for merging Java API into master

* Changed constructors to package private instead of private
* adding unit test for MKLDNN FullyConnected operator

* removing mkldnn filter

* removing mkldnn filter
* fix.

* add test.

* retrigger
* Addressed doc issues

* Update optimizer.py
…structions (apache#13267)

* Added command line alternative to IntelliJ

* Removed the duplicate file

* Fixed typos

* Fixed minor command issue
* fixing gradcam

* changed loading parameters code

* fixing type conversions issue with previous versions of matplotlib

* gradcam consolidation

* creating directory structures in utils

* changing location

* empty commit
* Adding info_gan example

* adjust paths of filenames

* Update index.md

* Update index.md

* Update index.md

* Update info_gan.md

Added an image

* Update info_gan.md

Applied some fixes

* Update info_gan.md

Applied some fixes

* Update info_gan.md

Applied some fixes

* Update info_gan.md

* Updated index.md file

* Updated index.md file

* change links

* Fixed typo

* Delete Untitled.ipynb

* Adding Vishaals comments

* Adding Anirudh's comments

* Fixed some bugs

* Adding Anirudh's comments

* some minor fixes
)

* Added Java SSD Inference Tutorial for website

* Added whitelisting to SSD tutorial

* Address PR feedback

* Marking intelliJ as optional
* add initial commit

* push back predictor

* name fix and bug fix

* update readme and script to run

* minor fix

* minor fix

* fix on doc

* update predictor
* Fix Sphinx python docstring formatting error (apache#13021).

Fixes apache#13021

* Update src/operator/nn/batch_norm.cc

Co-Authored-By: frankfliu <[email protected]>
srochel and others added 29 commits December 20, 2018 05:59
* merge NEWS.md from 1.4.x to master

* NEWS.md backport from v1.4.x to master
…e#13599)

* fallback to dense version for grad(reshape), grad(expand_dims)

* add _backward_reshape gpu version

* reshape test case comments

* fix gpu test

* remove mkldnn support for _backward_reshape
* Add Flatten before Gemm

* ONNX export test: Allow multiple inputs in forward pass

* ONNX export: Test for fully connected
…che#12977)

* Adding cpp-package directory to the Doxyfile. Updating the index.md file in c++ api directory.

* Updating the link to classes in C++ API to point to correct html file.

* Updated the links to use relative paths.

* Removed the extra slash character in the url

* Excluded the 3rdparty folder as per the review comment.
* Added timeout/retry (linear backoff) to docker cache download

* Units changed, as time.sleep takes seconds as argument

* Improved error handling

* Using retry decorator

* Added retry decorator to _login_dockerhub method

* Fixed wrong import
* Update test_gluon_trainer.py

* Update test_gluon_trainer.py

* test
* fp16 dot

* update mshadow

* update mshadow

* update mshadow
* fix the quantization script to support python2

* Fix comments, fix similiar issue in imagenet_inference.py
* ONNX test code cleanup

* Make tests use the common test case list

* Remove import test_cases

* Make Gluon backend rep common

* Partially enable broadcast tests

* Common function to populate tests

* Make backend common

* test models

* Test nodes

* ONNX export: Test for fully connected

* Edit CI scripts mxnet export test cleanup

* Further cleanup backend tests

* README

* Some corrections

* test case format for test_models
* Port of scala infer package to clojure

* Add inference examples

* Fix project.clj

* Update code for integration tests

* Address comments and add unit tests

* Add specs and simplify interface

* Minor nit

* Update README
…3728)

* tests

* remove optimizer and move op to contrib

* rename parameter
…e#13654)

* Logsoftmax, missing tests

* Support multiple outputs in Gluon backendrep

* Remove repeated unsqueeze test

* Allow multiple output support
* Common test caller

* Remove incorrect comment

* Make corrections to CI

* fix ci script
* ONNX import: Hardmax

* Fix lint errors

* add github link for issue with reshape
* add clarification for param_dict

* more tests for dist kvstore

* more unittests

* fix a bug

* more dist exception test

* revert optimizer list

* fix bug and comment

* fix doc rendering and lint

* add invalid sched test

* fix website

* trigger

* update doc
* Reorder module import orders for dist-kvstore

* more code comments
* Allow CMake based installation of cpp-package

* Add installation of missing nnvm headers

* Add documentation as to where public headers will be installed
https://github.com/apache/incubator-mxnet/blob/a38278ddebfcc9459d64237086cd7977ec20c70e/example/image-classification/train_imagenet.py#L42

When I try to train imagenet with this line commented, the train-accuracy reaches 99% while the validation-accuracy is only less than 50% (single machine, 8 GPUs, global batchsize=2048, Resnet50). Absolutely this is overfitting.

Then I uncomment this line and try again with the same experiment settings. This time both train and validation accuracy converge to about 70%. 

Thus, it seems that this data augmentation is pretty important for ImageNet training. Perhaps it will be better to uncomment this as default, so that future developers won't get confused by the over-fit issue.
@ymjiang ymjiang merged commit 255737c into master Jun 9, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.