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.

Sign up for GitHub

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

Jump to bottom

AutoAugment image augmentation #41

Closed
innat opened this issue Jan 12, 2022 · 9 comments
Closed

AutoAugment image augmentation #41

innat opened this issue Jan 12, 2022 · 9 comments
Labels
preprocessing roadmap Components on the roadmap. stat:contributions welcome

Comments

@innat
Copy link
Contributor

innat commented Jan 12, 2022

Paper: https://arxiv.org/abs/1805.09501 Cited by 832
Ref code. https://github.com/tensorflow/models/blob/ded32f0500604928e52e27fd3f678e694e5133b7/official/vision/image_classification/augment.py#L731
@LukeWood
Copy link
Contributor

I've been weighing adding this layer.

Thoughts on what it adds over RandAugment?

It looks as though it's much more customizable, which could be good to have.

@innat

@bhack
Copy link
Contributor

bhack commented Jan 28, 2022

I think that we could also evaluate differentiable augmentations like:

https://arxiv.org/abs/2104.04282
https://arxiv.org/abs/2109.15273
https://openreview.net/forum?id=St-53J9ZARf

@innat
Copy link
Contributor Author

innat commented Jan 28, 2022

(AFAIK), the advantages of auto-augment over rand-augment also bring computational costs. I think both of them are used and maintained in research, so it's ok to have them both.

@bhack bhack mentioned this issue Feb 7, 2022
Closed
@bhack
Copy link
Contributor

bhack commented Feb 8, 2022

I think that we could also evaluate differentiable augmentations like:
https://arxiv.org/abs/2104.04282
https://arxiv.org/abs/2109.15273
https://openreview.net/forum?id=St-53J9ZARf

/cc @haifeng-jin

@LukeWood
Copy link
Contributor

this should probably use RandomAugmentationPipeline.

@bhack
Copy link
Contributor

bhack commented May 3, 2022

I think that we could also evaluate differentiable augmentations like:
.....
https://openreview.net/forum?id=St-53J9ZARf

https://github.com/MSU-MLSys-Lab/DeepAA

@bhack bhack mentioned this issue May 5, 2022
Closed
@LukeWood LukeWood added the roadmap Components on the roadmap. label May 31, 2022
@LukeWood
Copy link
Contributor

I've received numerous inquiries about AutoAugment, so I definitely think we want this component.

@bhack
Copy link
Contributor

bhack commented Oct 21, 2022

Has #900 was cloud I will unify the references on this ticket.

@LukeWood
Copy link
Contributor

LukeWood commented May 4, 2023

Lets deprioritize in favor of RandAugment; which serves a similar purpose but is much easier to tune.

@LukeWood LukeWood closed this as completed May 4, 2023
freedomtan pushed a commit to freedomtan/keras-cv that referenced this issue Jul 20, 2023
@ariG23498
Add binary accuracy (keras-team#41)
070f2b8 
* chore: adding binary accuracy

* chore: fix docstring
freedomtan pushed a commit to freedomtan/keras-cv that referenced this issue Jul 20, 2023
@nkovela1 @fchollet
@ariG23498 @sampathweb @chenmoneygithub @haifeng-jin @grasskin
Adds RMSprop and AdamW optimizers with associated tests. (keras-team#59)
b49e12e 
* Add golden correctness tests for Adam and SGD

* Fix dtype issues

* Sync with main (keras-team#56)

* Minor touch ups

* Fix a pretty major bug

* Format code

* Big rethink of Variable API

* Make build-by-run the default build(), leveraging new zero_history KerasTensor mode

* Minor fixes

* Format code

* Switch back to build-by-eager-run for simplicity

* Add raise upon build failure

* Work around JAX bug.

* Add a few more tests.

* Add saving tests

* Adds test suite for SGD and golden correctness tests for all optimizers (keras-team#40)

* Add golden correctness tests for Adam and SGD

* Fix dtype issues

* Add binary accuracy (keras-team#41)

* chore: adding binary accuracy

* chore: fix docstring

* Add tests for add_loss and activity regularization.

* Reformat code

* Add ActivityRegularization layer

* Fix JAX CI.

* Add Lambda Callback (keras-team#42)

* Add LambdaCallback

* Add Lambda Callback

* Add Lambda Callback

* Rename lambda_callback_test.py

* Add einsum (keras-team#43)

* Add einsum

* address comments

* Fix format line length (keras-team#45)

* Add Embedding layer

* Shorten lines

* Add .vscode to .gitignore (keras-team#46)

* rm vscode settings

* add .vscode to gitignore

* Set demo program backend (keras-team#48)

* Add tests for training arg resolution in Layer.

* Implement mixed precision.

* Replace backend.execute with backend.numpy.XXX (keras-team#50)

* Add cosine similarity loss and update l2_normalize from regularizers (keras-team#34)

* Begin cosine loss

* Add testing for cosine similarity

* Fix formatting

* Docstring standardization

* Formatting

* Create numerical_utils

* Fix issue with call context lingering.

* Add the EarlyStopping callback (keras-team#44)

* add earlystopping callback

* addressing comments

* address comments

* addressing comments

* remove unused imports

* re-enable imports checks (keras-team#51)

* Add nn.one_hot (keras-team#52)

* Add GaussianDropout layer.

* Add GaussianNoise layer

* Add Categorical Accuracy Metric (keras-team#47)

* chore: adding categorical accuracy metric

* chore: reformat docstrings

* chore: reformat

* chore: ndims with len

* refactor the docstring

* Fix typos

* Implement masking.

---------

Co-authored-by: Francois Chollet <[email protected]>
Co-authored-by: Aritra Roy Gosthipaty <[email protected]>
Co-authored-by: Ramesh Sampath <[email protected]>
Co-authored-by: Chen Qian <[email protected]>
Co-authored-by: Haifeng Jin <[email protected]>
Co-authored-by: Gabriel Rasskin <[email protected]>

* Adds rmsprop optimizer and tests

* Add AdamW optimizer and tests, minor formatting changes

* Implemented formatting fixes

---------

Co-authored-by: Francois Chollet <[email protected]>
Co-authored-by: Aritra Roy Gosthipaty <[email protected]>
Co-authored-by: Ramesh Sampath <[email protected]>
Co-authored-by: Chen Qian <[email protected]>
Co-authored-by: Haifeng Jin <[email protected]>
Co-authored-by: Gabriel Rasskin <[email protected]>
freedomtan pushed a commit to freedomtan/keras-cv that referenced this issue Jul 20, 2023
@nkovela1 @fchollet
@ariG23498 @sampathweb @chenmoneygithub @haifeng-jin @grasskin
Adds Adagrad and Adadelta optimizers and associated tests. (keras-tea…
047485b 
…m#72)

* Add golden correctness tests for Adam and SGD

* Fix dtype issues

* Sync with main (keras-team#56)

* Minor touch ups

* Fix a pretty major bug

* Format code

* Big rethink of Variable API

* Make build-by-run the default build(), leveraging new zero_history KerasTensor mode

* Minor fixes

* Format code

* Switch back to build-by-eager-run for simplicity

* Add raise upon build failure

* Work around JAX bug.

* Add a few more tests.

* Add saving tests

* Adds test suite for SGD and golden correctness tests for all optimizers (keras-team#40)

* Add golden correctness tests for Adam and SGD

* Fix dtype issues

* Add binary accuracy (keras-team#41)

* chore: adding binary accuracy

* chore: fix docstring

* Add tests for add_loss and activity regularization.

* Reformat code

* Add ActivityRegularization layer

* Fix JAX CI.

* Add Lambda Callback (keras-team#42)

* Add LambdaCallback

* Add Lambda Callback

* Add Lambda Callback

* Rename lambda_callback_test.py

* Add einsum (keras-team#43)

* Add einsum

* address comments

* Fix format line length (keras-team#45)

* Add Embedding layer

* Shorten lines

* Add .vscode to .gitignore (keras-team#46)

* rm vscode settings

* add .vscode to gitignore

* Set demo program backend (keras-team#48)

* Add tests for training arg resolution in Layer.

* Implement mixed precision.

* Replace backend.execute with backend.numpy.XXX (keras-team#50)

* Add cosine similarity loss and update l2_normalize from regularizers (keras-team#34)

* Begin cosine loss

* Add testing for cosine similarity

* Fix formatting

* Docstring standardization

* Formatting

* Create numerical_utils

* Fix issue with call context lingering.

* Add the EarlyStopping callback (keras-team#44)

* add earlystopping callback

* addressing comments

* address comments

* addressing comments

* remove unused imports

* re-enable imports checks (keras-team#51)

* Add nn.one_hot (keras-team#52)

* Add GaussianDropout layer.

* Add GaussianNoise layer

* Add Categorical Accuracy Metric (keras-team#47)

* chore: adding categorical accuracy metric

* chore: reformat docstrings

* chore: reformat

* chore: ndims with len

* refactor the docstring

* Fix typos

* Implement masking.

---------

Co-authored-by: Francois Chollet <[email protected]>
Co-authored-by: Aritra Roy Gosthipaty <[email protected]>
Co-authored-by: Ramesh Sampath <[email protected]>
Co-authored-by: Chen Qian <[email protected]>
Co-authored-by: Haifeng Jin <[email protected]>
Co-authored-by: Gabriel Rasskin <[email protected]>

* Adds rmsprop optimizer and tests

* Add AdamW optimizer and tests, minor formatting changes

* Implemented formatting fixes

* Adds clip norm and clip value tests to Adam

* Adds Adagrad and Adadelta optimizers

* Applies fixes to formatting and deletes unnecessary kwargs

---------

Co-authored-by: Francois Chollet <[email protected]>
Co-authored-by: Aritra Roy Gosthipaty <[email protected]>
Co-authored-by: Ramesh Sampath <[email protected]>
Co-authored-by: Chen Qian <[email protected]>
Co-authored-by: Haifeng Jin <[email protected]>
Co-authored-by: Gabriel Rasskin <[email protected]>
freedomtan pushed a commit to freedomtan/keras-cv that referenced this issue Jul 20, 2023
@nkovela1 @fchollet
@ariG23498 @sampathweb @chenmoneygithub @haifeng-jin @grasskin
Adds all remaining Keras optimizers (Adamax, Adafactor, Nadam, and Ft…
3cf0a83 
…rl) (keras-team#80)

* Add golden correctness tests for Adam and SGD

* Fix dtype issues

* Sync with main (keras-team#56)

* Minor touch ups

* Fix a pretty major bug

* Format code

* Big rethink of Variable API

* Make build-by-run the default build(), leveraging new zero_history KerasTensor mode

* Minor fixes

* Format code

* Switch back to build-by-eager-run for simplicity

* Add raise upon build failure

* Work around JAX bug.

* Add a few more tests.

* Add saving tests

* Adds test suite for SGD and golden correctness tests for all optimizers (keras-team#40)

* Add golden correctness tests for Adam and SGD

* Fix dtype issues

* Add binary accuracy (keras-team#41)

* chore: adding binary accuracy

* chore: fix docstring

* Add tests for add_loss and activity regularization.

* Reformat code

* Add ActivityRegularization layer

* Fix JAX CI.

* Add Lambda Callback (keras-team#42)

* Add LambdaCallback

* Add Lambda Callback

* Add Lambda Callback

* Rename lambda_callback_test.py

* Add einsum (keras-team#43)

* Add einsum

* address comments

* Fix format line length (keras-team#45)

* Add Embedding layer

* Shorten lines

* Add .vscode to .gitignore (keras-team#46)

* rm vscode settings

* add .vscode to gitignore

* Set demo program backend (keras-team#48)

* Add tests for training arg resolution in Layer.

* Implement mixed precision.

* Replace backend.execute with backend.numpy.XXX (keras-team#50)

* Add cosine similarity loss and update l2_normalize from regularizers (keras-team#34)

* Begin cosine loss

* Add testing for cosine similarity

* Fix formatting

* Docstring standardization

* Formatting

* Create numerical_utils

* Fix issue with call context lingering.

* Add the EarlyStopping callback (keras-team#44)

* add earlystopping callback

* addressing comments

* address comments

* addressing comments

* remove unused imports

* re-enable imports checks (keras-team#51)

* Add nn.one_hot (keras-team#52)

* Add GaussianDropout layer.

* Add GaussianNoise layer

* Add Categorical Accuracy Metric (keras-team#47)

* chore: adding categorical accuracy metric

* chore: reformat docstrings

* chore: reformat

* chore: ndims with len

* refactor the docstring

* Fix typos

* Implement masking.

---------

Co-authored-by: Francois Chollet <[email protected]>
Co-authored-by: Aritra Roy Gosthipaty <[email protected]>
Co-authored-by: Ramesh Sampath <[email protected]>
Co-authored-by: Chen Qian <[email protected]>
Co-authored-by: Haifeng Jin <[email protected]>
Co-authored-by: Gabriel Rasskin <[email protected]>

* Adds rmsprop optimizer and tests

* Add AdamW optimizer and tests, minor formatting changes

* Implemented formatting fixes

* Adds clip norm and clip value tests to Adam

* Adds Adagrad and Adadelta optimizers

* Applies fixes to formatting and deletes unnecessary kwargs

* Adds Adamax and Adafactor and associated tests

* Adds Nadam and Ftrl optimizers and associated tests

---------

Co-authored-by: Francois Chollet <[email protected]>
Co-authored-by: Aritra Roy Gosthipaty <[email protected]>
Co-authored-by: Ramesh Sampath <[email protected]>
Co-authored-by: Chen Qian <[email protected]>
Co-authored-by: Haifeng Jin <[email protected]>
Co-authored-by: Gabriel Rasskin <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
preprocessing roadmap Components on the roadmap. stat:contributions welcome
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@bhack @LukeWood @innat