18 Pull requests merged by 8 people

• Fix _can_use_flash_attention.

  #21512 merged Jul 25, 2025

• fix torch module wrapper serialization error

  #21505 merged Jul 25, 2025

• fix TAP test

  #21515 merged Jul 25, 2025

• Keras <> NNX integration

  #21252 merged Jul 25, 2025

• [OpenVINO backend] add supporting for lists and tuples in get_ov_output

  #21495 merged Jul 24, 2025

• [OpenVINO backend] fix_transpose

  #21496 merged Jul 24, 2025

• [OpenVINO backend] fix numpy conversions

  #21498 merged Jul 24, 2025

• [OpenVINO backend] support export model from the supported backends to openvino format

  #21486 merged Jul 24, 2025

• Add Grain integration.

  #21494 merged Jul 24, 2025

• [OpenVINO backend] support categorical

  #21437 merged Jul 24, 2025

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

  #21408 merged Jul 24, 2025

• mlx - more changes to tests to pass CI

  #21415 merged Jul 24, 2025

• Update compute_output_spec method to support cbrt operation

  #21490 merged Jul 24, 2025

• docs: enhance clarity and consistency in README by correcting platform capitalization

  #21493 merged Jul 24, 2025

• Docs - Fixed Syntax Error in vectorized_map function example

  #21503 merged Jul 24, 2025

• Support shape -1 for slice op in the jax backend

  #21501 merged Jul 23, 2025

• Fix RandAugment behavior in tf graph mode

  #21499 merged Jul 22, 2025

• [OpenVINO backend] support take_along_axis

  #21447 merged Jul 21, 2025

4 Pull requests opened by 3 people

• Simulated OpenVINO Backend for Testing Unmerged PR Features with Memory Profiling

  #21500 opened Jul 22, 2025

• fix quantization save and load error

  #21504 opened Jul 23, 2025

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

  #21506 opened Jul 24, 2025

• Fix Discretization layer graph mode bug

  #21514 opened Jul 25, 2025

4 Issues closed by 4 people

• TorchModuleWrapper default `get_config` creates non serializable objects

  #19226 closed Jul 25, 2025

• Bug for _can_use_flash_attention

  #21507 closed Jul 25, 2025

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

  #21172 closed Jul 24, 2025

• [bug] keras issue on tpu-vm

  #19448 closed Jul 21, 2025

5 Issues opened by 5 people

• TypeError with single-output dictionary targets - "Expected float32, but got my_output_0 of type 'str'"

  #21513 opened Jul 25, 2025

• Specifying ONNX opset with export method

  #21511 opened Jul 25, 2025

• Error in densify_boxes?

  #21510 opened Jul 25, 2025

• Add daily CPU tests which also test with the latest Python version

  #21508 opened Jul 24, 2025

• keras.layers.Flatten() not updating shape information after transformation

  #21502 opened Jul 23, 2025

17 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 Jul 23, 2025 • 3 new comments

• Fix: UpSampling2D bilinear set_image_data_format(channels_first) bug

  #21456 commented on Jul 24, 2025 • 0 new comments

• Running TPU tests on linux-x86-ct6e-44-1tpu

  #21425 commented on Jul 23, 2025 • 0 new comments

• IGNORE THIS PR

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

• Fix stateful RNN to raise ValueError on batch size mismatch

  #21249 commented on Jul 25, 2025 • 0 new comments

• Add memory usage monitor callback

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

• Fix TorchModuleWrapper serialization issue

  #20869 commented on Jul 23, 2025 • 0 new comments

• ZeroDivisionError: integer modulo by zero | multi-device setup

  #21167 commented on Jul 25, 2025 • 0 new comments

• Cannot properly save the model using TorchModuleWrapper and load the model

  #21350 commented on Jul 24, 2025 • 0 new comments

• How does Keras MultiHeadAttention handle input feature dimension vs heads and reshape the output?

  #21219 commented on Jul 23, 2025 • 0 new comments

• Integrate Grain into Keras

  #21479 commented on Jul 23, 2025 • 0 new comments

• RematScope with mixed precision: autocast on kernels doesn't seem to trigger on the backwards pass

  #21448 commented on Jul 23, 2025 • 0 new comments

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

  #21487 commented on Jul 22, 2025 • 0 new comments

• Keras Fails to load quantized model

  #21378 commented on Jul 22, 2025 • 0 new comments

• MaP always 0. Using Yolov8

  #18935 commented on Jul 22, 2025 • 0 new comments

• evaluate() shows incorrect metric values in progress bar vs returned results

  #21301 commented on Jul 21, 2025 • 0 new comments

• 🚀OpenVINO Backend: GFIs for Further Development🚀

  #20910 commented on Jul 20, 2025 • 0 new comments