mlx - merge master into mlx #19657

lkarthee · 2024-05-02T04:17:25Z

No description provided.

* Add float8 training support * Add tests for fp8 training * Add `quantize_and_dequantize` test * Fix bugs and add float8 correctness tests * Cleanup * Address comments and cleanup * Add docstrings and some minor refactoring * Add `QuantizedFloat8DTypePolicy` * Add dtype policy setter * Fix torch dynamo issue by using `self._dtype_policy` * Improve test coverage

* Add LoRA to `BaseConv` * Add tests * Fix typo * Fix tests * Fix tests

* feat(losses): add Tversky loss implementation * adjusted documentation

…9522) - added `Layer.metrics` to return all metrics owned by the layer and its sub-layers recursively. - `Layer.metrics_variables` now returns variables from all metrics recursively, not just the layer and its direct sub-layers. - `Model.metrics` now returns all metrics recursively, not just the model level metrics. - `Model.metrics_variables` now returns variables from all metrics recursively, not just the model level metrics. - added test coverage to test metrics and variables 2 levels deep. This is consistent with the Keras 2 behavior and how `Model/Layer.variables` and `Model/Layer.weights` work.

…nd Test Coverage for `Preprocessing Layers` (keras-team#19513) * Add tests for CategoryEncoding class in category_encoding_test.py * fix * Fix IndexLookup class initialization and add test cases * Add test case for IndexLookupLayerTest without vocabulary * Fix IndexLookup class initialization * Add normalization test cases * Add test cases for Hashing class * Fix value range validation error in RandomBrightness class * Refactor IndexLookup class initialization and add test cases * Reffix ndexLookup class initialization and afix est cases

* API Generator for Keras * API Generator for Keras * Generates API Gen via api_gen.sh * Remove recursive import of _tf_keras * Generate API Files via api_gen.sh

…eras-team#19529) This works the same way as in Keras 2, whereby the metrics are returned directly from the logs if the set of keys doesn't match the model metrics.

* Use temp dir and abs path * Use temp dir and abs path * Update Readme

… float8 training (keras-team#19534) * Fix gradient accumulation with `overwrite_with_gradient` in float8 training * Add comments * Fix annotation

* Add logic to prevent double quantization * Add detailed info for double quantization error * Update error msg

…eam#19635)

* Fix greedy ctc decode * Remove print * Fix `tf.nn.ctc_beam_search_decoder` * Change default `mask_index` to `0` * Fix losses test * Update

…#19636) * Ensure the same rule applies for np arrays in autocasting * Trigger CI by adding docstring * Update * Update docstring

…as-team#19594) * Fix and test math functions for jax backend * run /workspaces/keras/shell/format.sh * refix * fix * fix _get_complex_tensor_from_tuple * fix * refix * Fix istft function to handle inputs with less than 2 dimensions * fix * Fix ValueError in istft function for inputs with less than 2 dimensions

…9640) With Torch, `x.shape` returns a `torch.Size`, which is a subclass of `tuple` but can cause different behaviors. In particular `convert_to_tensor` does not work on `torch.Size`. This fixes keras-team#18900

* Enable cudnn rnns when dropout is set * Fix

Bumps the github-actions group with 2 updates: [actions/upload-artifact](https://github.com/actions/upload-artifact) and [github/codeql-action](https://github.com/github/codeql-action). Updates `actions/upload-artifact` from 4.3.1 to 4.3.3 - [Release notes](https://github.com/actions/upload-artifact/releases) - [Commits](actions/upload-artifact@5d5d22a...6546280) Updates `github/codeql-action` from 3.24.9 to 3.25.3 - [Release notes](https://github.com/github/codeql-action/releases) - [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md) - [Commits](github/codeql-action@1b1aada...d39d31e) --- updated-dependencies: - dependency-name: actions/upload-artifact dependency-type: direct:production update-type: version-update:semver-patch dependency-group: github-actions - dependency-name: github/codeql-action dependency-type: direct:production update-type: version-update:semver-minor dependency-group: github-actions ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Bumps the python group with 2 updates: torch and torchvision. Updates `torch` from 2.2.1+cu121 to 2.3.0+cu121 Updates `torchvision` from 0.17.1+cu121 to 0.18.0+cu121 --- updated-dependencies: - dependency-name: torch dependency-type: direct:production update-type: version-update:semver-minor dependency-group: python - dependency-name: torchvision dependency-type: direct:production update-type: version-update:semver-minor dependency-group: python ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

…as-team#19655) This reverts commit 09133f4.

codecov-commenter · 2024-05-02T04:23:06Z

Codecov Report

Attention: Patch coverage is 88.15713% with 205 lines in your changes are missing coverage. Please review.

❗ No coverage uploaded for pull request base (mlx@4c90dfb). Click here to learn what that means.

Files	Patch %	Lines
keras/src/backend/common/backend_utils.py	59.64%	32 Missing and 14 partials ⚠️
keras/src/backend/tensorflow/numpy.py	86.75%	23 Missing and 15 partials ⚠️
keras/src/backend/numpy/nn.py	94.52%	5 Missing and 7 partials ⚠️
keras/src/saving/saving_lib.py	80.00%	6 Missing and 6 partials ⚠️
...s/src/trainers/data_adapters/py_dataset_adapter.py	72.50%	9 Missing and 2 partials ⚠️
keras/src/ops/numpy.py	87.65%	6 Missing and 4 partials ⚠️
keras/src/backend/jax/nn.py	95.78%	3 Missing and 5 partials ⚠️
keras/src/utils/model_visualization.py	0.00%	8 Missing ⚠️
keras/api/_tf_keras/keras/ops/__init__.py	0.00%	6 Missing ⚠️
keras/src/backend/tensorflow/nn.py	90.47%	2 Missing and 4 partials ⚠️
... and 18 more

Additional details and impacted files

@@          Coverage Diff           @@
##             mlx   #19657   +/-   ##
======================================
  Coverage       ?   75.56%           
======================================
  Files          ?      508           
  Lines          ?    47212           
  Branches       ?     8681           
======================================
  Hits           ?    35675           
  Misses         ?     9795           
  Partials       ?     1742

Flag	Coverage Δ
keras	`75.41% <88.15%> (?)`
keras-jax	`59.78% <50.54%> (?)`
keras-numpy	`54.33% <46.73%> (?)`
keras-tensorflow	`61.08% <58.69%> (?)`
keras-torch	`59.80% <47.94%> (?)`

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

james77777778 and others added 30 commits April 15, 2024 10:06

Add LoRA to ConvND layers (keras-team#19516)

38a0caa

* Add LoRA to `BaseConv` * Add tests * Fix typo * Fix tests * Fix tests

Add path to run keras on dm-tree when optree is not available.

c458201

feat(losses): add Tversky loss implementation (keras-team#19511)

61bbff5

* feat(losses): add Tversky loss implementation * adjusted documentation

Update KLD docs

6503b6d

Update IoU ignore_class handling

2ff3f94

Merge branch 'master' of github.com:keras-team/keras

fed28a7

Add test for spectral norm

d33b585

Add missing test decorator

a4e2716

Fix torch test

344b3e3

Fix code format

559f1dd

Generate API (keras-team#19530)

1937d48

* API Generator for Keras * API Generator for Keras * Generates API Gen via api_gen.sh * Remove recursive import of _tf_keras * Generate API Files via api_gen.sh

Update APIs

9fd057b

Added metrics from custom train_step/test_step are now returned. (k…

e57b138

…eras-team#19529) This works the same way as in Keras 2, whereby the metrics are returned directly from the logs if the set of keys doesn't match the model metrics.

Use temp dir and abs path in api_gen.py (keras-team#19533)

13cb10d

* Use temp dir and abs path * Use temp dir and abs path * Update Readme

Update API

9e64058

Merge branch 'master' of github.com:keras-team/keras

4d0a5de

Fix gradient accumulation when using overwrite_with_gradient during…

4adfbd4

… float8 training (keras-team#19534) * Fix gradient accumulation with `overwrite_with_gradient` in float8 training * Add comments * Fix annotation

Update code path in ignore path (keras-team#19537)

47b5ba5

Add operations per run (keras-team#19538)

1c9e01c

Include input shapes in model visualization.

69f3abd

Merge branch 'master' of github.com:keras-team/keras

7386704

Add pad_to_aspect_ratio feature in ops.image.resize

1228acb

Add pad_to_aspect_ratio feature in Resizing layer.

7cd5c60

Fix incorrect usage of quantize (keras-team#19541)

b955665

* Add logic to prevent double quantization * Add detailed info for double quantization error * Update error msg

Add eigh op.

ab017ea

Add keepdim in argmax/argmin.

65d0e09

Fix small bug in model.save_weights (keras-team#19545)

1135431

lpizzinidev and others added 14 commits April 28, 2024 11:39

fix(ops): specify NonZero output dtype and add test coverage (keras-t…

d5c9540

…eam#19635)

Fix ops.ctc_decode (keras-team#19633)

880f0cd

* Fix greedy ctc decode * Remove print * Fix `tf.nn.ctc_beam_search_decoder` * Change default `mask_index` to `0` * Fix losses test * Update

Ensure the same rule applies for np arrays in autocasting (keras-team…

4cb5671

…#19636) * Ensure the same rule applies for np arrays in autocasting * Trigger CI by adding docstring * Update * Update docstring

Return a tuple from ops.shape with the Torch backend. (keras-team#1…

9f4da51

…9640) With Torch, `x.shape` returns a `torch.Size`, which is a subclass of `tuple` but can cause different behaviors. In particular `convert_to_tensor` does not work on `torch.Size`. This fixes keras-team#18900

support conv3d on cpu for TF (keras-team#19641)

f01e99a

Enable cudnn rnns when dropout is set (keras-team#19645)

1b8f7b7

* Enable cudnn rnns when dropout is set * Fix

Fix plot_model for input dicts.

e85c12b

Merge branch 'master' of github.com:keras-team/keras

751db1b

Fix deprecation warning in torch

9a69ece

Revert "Bump the python group with 2 updates (keras-team#19654)" (ker…

1941c30

…as-team#19655) This reverts commit 09133f4.

Merge remote-tracking branch 'upstream/master' into mlx

92f71b4

google-ml-butler bot added the size:XL label May 2, 2024

google-ml-butler bot assigned gbaned May 2, 2024

gbaned requested a review from fchollet May 3, 2024 03:28

google-ml-butler bot added the awaiting review label May 3, 2024

fchollet approved these changes May 3, 2024

View reviewed changes

google-ml-butler bot added kokoro:force-run ready to pull Ready to be merged into the codebase labels May 3, 2024

fchollet merged commit 5a3542b into keras-team:mlx May 3, 2024
6 of 8 checks passed

google-ml-butler bot removed awaiting review ready to pull Ready to be merged into the codebase kokoro:force-run labels May 3, 2024

lkarthee deleted the merge_upstream_master branch May 3, 2024 17:28

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

mlx - merge master into mlx #19657

mlx - merge master into mlx #19657

lkarthee commented May 2, 2024

codecov-commenter commented May 2, 2024 •

edited

Loading

mlx - merge master into mlx #19657

mlx - merge master into mlx #19657

Conversation

lkarthee commented May 2, 2024

codecov-commenter commented May 2, 2024 • edited Loading

Codecov Report

codecov-commenter commented May 2, 2024 •

edited

Loading