13 Pull requests merged by 9 people

• Fix nnx object state

  #21565 merged Aug 9, 2025

• [OpenVINO backend] Support numpy.deg2rad

  #21524 merged Aug 8, 2025

• Fix the Doc of alpha in mobilenet_v3

  #21563 merged Aug 8, 2025

• Fix mixed dtypes issue in nn.dot_product_attention.

  #21558 merged Aug 8, 2025

• It worth not hiding exceptions

  #21432 merged Aug 8, 2025

• docs: Keras 3 evaluate() and compiled metrics clarification

  #21506 merged Aug 8, 2025

• [OpenVINO backend] Remove usage of deprecated openvino.runtime

  #21418 merged Aug 8, 2025

• Add assume_unique and invert arguments to keras.ops.isin

  #21552 merged Aug 7, 2025

• Fix broken code block rendering in histogram documentation

  #21555 merged Aug 7, 2025

• Force computation and data placement on the CPU when using a Grain data pipeline.

  #21553 merged Aug 7, 2025

• [OpenVINO backend] support slice_update

  #21549 merged Aug 5, 2025

• Implement isneginf function in keras.ops

  #21538 merged Aug 4, 2025

• [OpenVINO backend] fix __getitem__ and convert_to_tensor issues

  #21545 merged Aug 4, 2025

9 Pull requests opened by 8 people

• Fix support for optional inputs in model.fit

  #21548 opened Aug 4, 2025

• feat(quantization): Add GPTQ n-bit quantization support

  #21551 opened Aug 6, 2025

• Added cholesky inverse operation to all the backends

  #21554 opened Aug 7, 2025

• feat: Add a predict_proba method on SKLearnClassifier

  #21556 opened Aug 7, 2025

• Refactor(JAX): Use jax.jit's out_shardings instead of _enforce_jax_state_sharding

  #21559 opened Aug 8, 2025

• Model pruning

  #21561 opened Aug 8, 2025

• Add daily Python 3.13 CPU-only tests to nightly workflow

  #21566 opened Aug 8, 2025

• [OpenVINO backend] Support numpy.prod

  #21567 opened Aug 9, 2025

• Proposed fix for issue #21519: Reshape layer does not handle -1 shape…

  #21568 opened Aug 9, 2025

6 Issues closed by 5 people

• Fix the Doc of `alpha` in `mobilenet_v3`

  #21562 closed Aug 8, 2025

• Trainer.evaluate and Trainer.metrics_names do not interact as specified in the docs

  #21487 closed Aug 8, 2025

• Specifying ONNX opset with export method

  #21511 closed Aug 7, 2025

• Training loop from scratch with JAX doesn't converge

  #21489 closed Aug 5, 2025

• MaP always 0. Using Yolov8

  #18935 closed Aug 5, 2025

• How to specify the device on which the model resides

  #21537 closed Aug 4, 2025

3 Issues opened by 3 people

• Mismatch on `output_mode` in func `IntegerLookup`

  #21564 opened Aug 8, 2025

• Adding a `predict_proba` method on SKLearnClassifier

  #21547 opened Aug 4, 2025

• get_config() of layers not more called in keras.Model if __init__ signature contains custom kwargs

  #21546 opened Aug 4, 2025

50 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.

• numpy.meshgrid implementation

  #21292 commented on Aug 8, 2025 • 4 new comments

• Fix stateful RNN to raise ValueError on batch size mismatch

  #21249 commented on Aug 8, 2025 • 3 new comments

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

  #21253 commented on Aug 8, 2025 • 2 new comments

• Resolve: [Good First Issue][Keras 3 OpenVINO Backend]: Support numpy.linspace operation #29485

  #21140 commented on Aug 8, 2025 • 1 new comment

• Fix TypeError in keras.ops.searchsorted by using shape instead of len

  #21472 commented on Aug 10, 2025 • 1 new comment

• [Documentation] Updated Categorical Focal Crossentropy Loss Docstring

  #21044 commented on Aug 8, 2025 • 1 new comment

• [OpenVino BackEnd] support np.diagnol for ov BackEnd

  #21156 commented on Aug 8, 2025 • 1 new comment

• OpenVINO: linspace

  #21209 commented on Aug 8, 2025 • 1 new comment

• Fix typos in additive_attention

  #20711 commented on Aug 8, 2025 • 1 new comment

• Fix typos in string_lookup

  #20667 commented on Aug 8, 2025 • 1 new comment

• 🗺️ Keras Development Roadmap

  #19519 commented on Aug 5, 2025 • 0 new comments

• Add support for numpy.flip in OpenVINO backend

  #21132 commented on Aug 8, 2025 • 0 new comments

• Support numpy.prod operation

  #21188 commented on Aug 8, 2025 • 0 new comments

• implementation of logaddexp in openvino backend (issue #29489)

  #21216 commented on Aug 8, 2025 • 0 new comments

• [OpenVINO Backend] Support for numpy.flip

  #21230 commented on Aug 8, 2025 • 0 new comments

• Add memory usage monitor callback

  #21245 commented on Aug 7, 2025 • 0 new comments

• This PR implements the numpy.logspace function for the OpenVINO backend as requested in issue #30114.

  #21266 commented on Aug 8, 2025 • 0 new comments

• Implement numpy.prod method for OpenVino backend

  #21282 commented on Aug 8, 2025 • 0 new comments

• use native dumpy conversion when available

  #21288 commented on Aug 8, 2025 • 0 new comments

• Added openvino backend support for numpy.median (Issue #30115)

  #21379 commented on Aug 8, 2025 • 0 new comments

• Openvino backend improvements

  #21381 commented on Aug 10, 2025 • 0 new comments

• Fix(OpenVINO): Ensure numpy.prod passes tests and handles boolean inp…

  #21403 commented on Aug 8, 2025 • 0 new comments

• [Keras 3 OpenVINO Backend]: Support numpy.prod operation

  #21405 commented on Aug 10, 2025 • 0 new comments

• [Keras 3 OpenVINO Backend]: Support numpy.flip operation

  #21410 commented on Aug 8, 2025 • 0 new comments

• [OpenVINO backend] Support numpy.logaddexp

  #21522 commented on Aug 8, 2025 • 0 new comments

• TESTING ONLY (TPU test)

  #21528 commented on Aug 6, 2025 • 0 new comments

• Add public register_call_context_args method to Layer

  #21543 commented on Aug 10, 2025 • 0 new comments

• [Feature Suggestion] Improving Built-In Visualization System With Layered-Style Architecture Generation

  #21421 commented on Aug 6, 2025 • 0 new comments

• Cannot load weight of pre-trained models

  #21529 commented on Aug 6, 2025 • 0 new comments

• Keras 3.7 Broke My Code

  #20568 commented on Aug 7, 2025 • 0 new comments

• Training performance degradation after switching from Keras 2 mode to Keras 3 using Tensorflow

  #20283 commented on Aug 7, 2025 • 0 new comments

• Tensorflow MultiWorkerMirroredStrategy and input dictionaries

  #20329 commented on Aug 7, 2025 • 0 new comments

• Keras Fails to load quantized model

  #21378 commented on Aug 8, 2025 • 0 new comments

• Reshape layer does not handle -1 shape infor dynamically

  #21519 commented on Aug 8, 2025 • 0 new comments

• Multi-backend support for timeseries_dataset_from_array()

  #21536 commented on Aug 8, 2025 • 0 new comments

• Custom call context args are not supported for the common methods for model calling (e.g. fit, predict, evaluate, ...)

  #21530 commented on Aug 8, 2025 • 0 new comments

• Model.quantize() increases memory allocated.

  #21518 commented on Aug 11, 2025 • 0 new comments

• Modified example in depthwise_conv1d file

  #20268 commented on Aug 8, 2025 • 0 new comments

• Fix gradient tape issue in test_gradient_tape_issue.py

  #20277 commented on Aug 8, 2025 • 0 new comments

• Added description for keras.constraints.serialize

  #20628 commented on Aug 8, 2025 • 0 new comments

• Fix to issue #20616

  #20683 commented on Aug 8, 2025 • 0 new comments

• fix tracking issue with lazy filled attributes

  #20783 commented on Aug 8, 2025 • 0 new comments

• Perceptual Similarity loss

  #20844 commented on Aug 8, 2025 • 0 new comments

• [Documentation] Added missing docstrings to 92 function(s) in examples.

  #20908 commented on Aug 8, 2025 • 0 new comments

• [OpenVINO Backend] Support np.diag

  #20967 commented on Aug 8, 2025 • 0 new comments

• 20954. adding a some changes line no.423

  #20991 commented on Aug 8, 2025 • 0 new comments

• Add numpy.logaddexp to openvino backend

  #21043 commented on Aug 8, 2025 • 0 new comments

• Updated numpy.flip using openvino opset

  #21057 commented on Aug 8, 2025 • 0 new comments

• [OpenVINO Backend] Support: Implemented numpy.diagonal for keras ov backend

  #21086 commented on Aug 8, 2025 • 0 new comments

• [OpenVINO Backend] Support: Implemented numpy.digitize for keras ov backend

  #21087 commented on Aug 8, 2025 • 0 new comments