10 Pull requests merged by 6 people

• Fix divide by zero error in math_test.py

  #21419 merged Jun 25, 2025

• Fix JAX GPU tests.

  #21402 merged Jun 24, 2025

• Disable loading functions within deserialization.

  #21412 merged Jun 24, 2025

• Add "format" support for latest JAX.

  #21406 merged Jun 20, 2025

• [OpenVINO Backend] update __getitem__

  #21359 merged Jun 20, 2025

• [OpenVINO Backend] support while_loop

  #21369 merged Jun 20, 2025

• Fix crash when sys.stdout.encoding is None

  #21392 merged Jun 18, 2025

• Remove references to jax.experimental.layout.Layout.

  #21400 merged Jun 18, 2025

• Fix missing and fragile scikit-learn imports in Keras sklearn wrappers

  #21387 merged Jun 18, 2025

• Add support for multiple dynamic dimensions to Flatten layer on TF.

  #21399 merged Jun 18, 2025

10 Pull requests opened by 7 people

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

  #21403 opened Jun 18, 2025

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

  #21405 opened Jun 19, 2025

• update a small feature in __getitem__ for openvino backend

  #21407 opened Jun 21, 2025

• [OpenVINO backend] support tri, triu, and tril

  #21408 opened Jun 21, 2025

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

  #21410 opened Jun 22, 2025

• [OpenVINO backend] supporting mul with bools

  #21411 opened Jun 22, 2025

• Improve error message for class_weight with PyTorch DataLoaders

  #21414 opened Jun 24, 2025

• `mlx` - more changes to tests to pass CI

  #21415 opened Jun 24, 2025

• Enable ROCm devices to correctly use keras.layers.CuDNNGRU and keras.…

  #21417 opened Jun 24, 2025

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

  #21418 opened Jun 24, 2025

11 Issues closed by 5 people

• loss_weights depending on epoch number

  #20294 closed Jun 25, 2025

• Is this Bug?

  #20984 closed Jun 25, 2025

• RNN constructor should use raise TypeError and not ValueError

  #21391 closed Jun 20, 2025

• Cannot generate heatmaps using Gradcam library

  #20475 closed Jun 20, 2025

• TimeDistributed layer computational time

  #21179 closed Jun 19, 2025

• Error exporting model with multiple outputs using model.export when specifying input_signature

  #21258 closed Jun 19, 2025

• keras.utils.timeseries_dataset_from_array bugs 1

  #21306 closed Jun 19, 2025

• AttributeError in SKLearnClassifier and SKLearnRegressor wrappers due to missing sklearn.utils.multiclass

  #21386 closed Jun 18, 2025

• Bug with `Flatten()` and TensorFlow backend

  #21380 closed Jun 18, 2025

• Function `openvino.core.custom_gradient()` should be a class

  #20953 closed Jun 18, 2025

• Documentation update guide required

  #21322 closed Jun 18, 2025

2 Issues opened by 2 people

• Weight loading: `skip_mismatch` is not respected for H5 format

  #21404 opened Jun 19, 2025

• keras.layers.UpSampling2D bilinear channel first bug

  #21401 opened Jun 18, 2025

23 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 Backend] support ops.slice_update

  #21362 commented on Jun 24, 2025 • 1 new comment

• IGNORE THIS PR

  #21389 commented on Jun 24, 2025 • 0 new comments

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

  #21379 commented on Jun 21, 2025 • 0 new comments

• Modified compute_output_shape function to handle broadcasting behavior in layers.Rescaling

  #21351 commented on Jun 23, 2025 • 0 new comments

• [OpenVINO backend] Update `ndim` behavior

  #21317 commented on Jun 24, 2025 • 0 new comments

• Implement numpy.prod method for OpenVino backend

  #21282 commented on Jun 24, 2025 • 0 new comments

• Keras <> NNX integration

  #21252 commented on Jun 20, 2025 • 0 new comments

• Add memory usage monitor callback

  #21245 commented on Jun 23, 2025 • 0 new comments

• Support numpy.prod operation

  #21188 commented on Jun 22, 2025 • 0 new comments

• MultiHeadAttention's use_causal_mask is broken

  #21284 commented on Jun 25, 2025 • 0 new comments

• Unable to see any speedup from multi-GPU training using JAX backend for LSTM / GRU

  #20998 commented on Jun 24, 2025 • 0 new comments

• Progress bar showing wrong step number

  #20861 commented on Jun 23, 2025 • 0 new comments

• Incorrect error message given for named inputs when unnamed input provided to fit()

  #19127 commented on Jun 23, 2025 • 0 new comments

• Documentation: better info about class_weight and/or DataAdapters

  #21355 commented on Jun 23, 2025 • 0 new comments

• Incorrect values returned by the evaluate method in Model API

  #20788 commented on Jun 23, 2025 • 0 new comments

• Data-Augmentation layers can't be exported to ONNX

  #21346 commented on Jun 21, 2025 • 0 new comments

• keras.layers.Normalization#adapt() does not support keras.utils.PyDataset as input.

  #21300 commented on Jun 20, 2025 • 0 new comments

• TF+XLA+Mixed Precision: Keras fails to compute gradients

  #20931 commented on Jun 19, 2025 • 0 new comments

• What is correct behavior of keras.ops.eye?

  #20616 commented on Jun 19, 2025 • 0 new comments

• Failed to convert LSTM layers to ONNX on Torch backend

  #21390 commented on Jun 18, 2025 • 0 new comments

• Custom Pooling Layer for CNN

  #20282 commented on Jun 18, 2025 • 0 new comments

• LossScaleOptimizer crash under jit_compile=True with MirroredStrategy and mixed_precision

  #21366 commented on Jun 18, 2025 • 0 new comments

• The documentation for Conv1DTranspose does not state that the CPU does not support dilation rates larger than 1

  #20408 commented on Jun 18, 2025 • 0 new comments