17 Pull requests merged by 12 people

• Fix take_along_axis with TensorFlow with non-int32 indices.

  #21256 merged May 7, 2025

• [Keras Ops] view_as_complex() and view_as_real()

  #21221 merged May 5, 2025

• Interleave optimizer variable creation to restore backward-compatibility

  #21247 merged May 5, 2025

• Add NotImplementedError for angle & bartlett on openvino

  #21246 merged May 5, 2025

• Implement blackman function in keras.ops

  #21235 merged May 4, 2025

• Make take_along_axis with TF backend compilable.

  #21239 merged May 4, 2025

• mlx - remat and continued test updates

  #21244 merged May 4, 2025

• Fix message formatting

  #21241 merged May 3, 2025

• Openvino numpy ravel

  #21231 merged May 1, 2025

• Bump github/codeql-action from 3.28.13 to 3.28.16 in the github-actions group

  #21237 merged May 1, 2025

• Fix stacked RNN with mask in JAX & Numpy backends

  #21224 merged May 1, 2025

• Implement bartlett function in keras.ops

  #21214 merged Apr 30, 2025

• Don't create unused optimizer variables.

  #21232 merged Apr 30, 2025

• Updated confusion_metrics.py

  #21227 merged Apr 30, 2025

• Adds _register_call_context_args to declare and use call-context arguments.

  #21222 merged Apr 30, 2025

• Update conftest.py

  #21220 merged Apr 30, 2025

• Clear static loss-scale for inner optimizer in LossScaleOptimizer.

  #21233 merged Apr 30, 2025

10 Pull requests opened by 10 people

• [OpenVINO Backend] Support for numpy.flip

  #21230 opened Apr 30, 2025

• Bump the python group with 2 updates

  #21236 opened May 1, 2025

• Bug fixes in `plot_model` with nested models.

  #21243 opened May 2, 2025

• Add memory usage monitor callback

  #21245 opened May 3, 2025

• Simplify the logic of `quantize` and float8 build.

  #21248 opened May 4, 2025

• Fix stateful RNN to raise ValueError on batch size mismatch

  #21249 opened May 4, 2025

• Keras <> NNX integration

  #21252 opened May 5, 2025

• [Keras 3 OpenVINO Backend]: Support numpy.logspace operation #30114

  #21253 opened May 6, 2025

• Fixed issue with dot_product_attention when using TPU.

  #21254 opened May 6, 2025

• Add openvino backend support for numpy.hstack

  #21257 opened May 7, 2025

8 Issues closed by 3 people

• Gemma quantize method not implemented?

  #21085 closed May 7, 2025

• Inconsistent `mode="auto"` behavior in `ModelCheckpoint` vs `EarlyStopping` and `ReduceLROnPlateau`

  #21118 closed May 7, 2025

• implement of muon optimizer

  #20952 closed May 5, 2025

• qwen3 tokenizer convert fail

  #21242 closed May 2, 2025

• Truncation during non-interactive model summary printing

  #21107 closed May 2, 2025

• Keras featurespace error in environment without TensoFlow

  #21009 closed May 1, 2025

• Support for type hints is not implemented

  #21021 closed May 1, 2025

• Improved Backend Checking

  #21005 closed Apr 30, 2025

4 Issues opened by 4 people

• StringLookup does not work on torch.Tensor

  #21255 opened May 6, 2025

• cannot find keras.model.layers definition anywhere on the keras doc

  #21240 opened May 2, 2025

• LayerNormalization with rms_scaling documentation is different from implementation

  #21234 opened Apr 30, 2025

• Keras 3 may not work with PyTorch DirectML

  #21228 opened Apr 30, 2025

27 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• OpenVINO: linspace

  #21209 commented on May 5, 2025 • 5 new comments

• DRAFT: Add custom variable updater.

  #21225 commented on Apr 30, 2025 • 2 new comments

• Add an easy way to run a script for a few steps only

  #21217 commented on May 1, 2025 • 0 new comments

• Added numpy.outer function using opset14

  #21213 commented on May 2, 2025 • 0 new comments

• Adding functional CompositeLayer

  #21099 commented on May 6, 2025 • 0 new comments

• Implement OpenVINO Decomposition for numpy.hstack Operation

  #21017 commented on May 1, 2025 • 0 new comments

• "PyDataset: init does not call super().init(**kwargs) causing workers, use_multiprocessing, and max_queue_size parameters to be ignored."

  #21068 commented on May 7, 2025 • 0 new comments

• custom_gradient not working with JAX backend

  #21105 commented on May 6, 2025 • 0 new comments

• Add `on_epoch_begin` to `utils.Sequence`

  #19749 commented on May 5, 2025 • 0 new comments

• MLX Backend

  #19571 commented on May 5, 2025 • 0 new comments

• 🚀OpenVINO Backend: GFIs for Further Development🚀

  #20910 commented on May 5, 2025 • 0 new comments

• old training model load error but new training is ok for colab

  #21103 commented on May 5, 2025 • 0 new comments

• Memory Usage Monitoring Callback

  #21150 commented on May 4, 2025 • 0 new comments

• stateful=True RNN silently broadcasts input when batch_size mismatch occurs

  #21183 commented on May 4, 2025 • 0 new comments

• Keras + JAX shape inference error

  #21155 commented on May 3, 2025 • 0 new comments

• `plot_model` does not work for all models in `keras.applications`

  #19717 commented on May 2, 2025 • 0 new comments

• SavedModel format doesn't support masking in Keras 3

  #21133 commented on May 1, 2025 • 0 new comments

• plot_model bugs with functional submodels

  #21119 commented on May 1, 2025 • 0 new comments

• Alias `KerasTensor` as `Tensor`

  #21210 commented on May 1, 2025 • 0 new comments

• ModelCheckpoint should throw an exception/provide warning if non-existing metric is being monitored

  #21109 commented on May 1, 2025 • 0 new comments

• Unexpected file structure when serializing a model with layer called summary_layer

  #21166 commented on May 1, 2025 • 0 new comments

• keras.ops.print: `tf.print`; `jax.debug.print`; …

  #21137 commented on May 1, 2025 • 0 new comments

• remat : The custom_gradient decorator currently supports keywords arguments only when eager execution is enabled.

  #21207 commented on Apr 30, 2025 • 0 new comments

• add RandomApply and RandomChoice

  #20727 commented on Apr 30, 2025 • 0 new comments

• TimeDistributed layer computational time

  #21179 commented on Apr 30, 2025 • 0 new comments

• Distributed training with tensorflow doesn't work with Keras 3.7 and above

  #21172 commented on Apr 30, 2025 • 0 new comments

• DTensor support in TF backend

  #18473 commented on Apr 30, 2025 • 0 new comments