9 Pull requests merged by 6 people

• Fix torch gpu CI

  #20696 merged Dec 28, 2024

• Add random_posterization processing layer

  #20688 merged Dec 27, 2024

• Introduces support for exporting SavedModel in the torch backend using torch-xla

  #20685 merged Dec 27, 2024

• minor: Fix Functional API guide

  #20694 merged Dec 27, 2024

• fix attention output with symbolic tensors and attention scores

  #20689 merged Dec 26, 2024

• Add random_color_degeneration processing layer

  #20679 merged Dec 25, 2024

• Fix: Return Attention Scores when return_attention_scores=True

  #20684 merged Dec 24, 2024

• Change the import order of pydot and its older forks

  #20682 merged Dec 23, 2024

• Add training status condition during image processing

  #20677 merged Dec 22, 2024

3 Pull requests opened by 3 people

• Fix issue with nested model output as input

  #20678 opened Dec 23, 2024

• Fix to issue #20616

  #20683 opened Dec 23, 2024

• Add random_sharpness processing layer

  #20697 opened Dec 28, 2024

10 Issues closed by 3 people

• ValueError: Only instances of keras.Layer can be added to a Sequential model. Received: <tensorflow_probability.python.layers.dense_variational_v2.DenseVariational object at 0x000001C67AF5B4D0> (of type <class 'tensorflow_probability.python.layers.dense_variational_v2.DenseVariational'>)

  #19894 closed Dec 27, 2024

• model.fit - class_weight broken

  #20542 closed Dec 27, 2024

• Cannot call AttentionLayer with symbolic tensors and return_attention_scores=True in Keras 3

  #20621 closed Dec 26, 2024

• Loading model fails: can only concatenate tuple

  #19776 closed Dec 26, 2024

• ValueError when trying to load a .keras model created using Functional API

  #19803 closed Dec 26, 2024

• InvalidArgumentError: Graph execution error:

  #20006 closed Dec 26, 2024

• Accuracy is lost after save_weights/load_weights

  #20524 closed Dec 26, 2024

• Need to explicitly specify `x` and `y` instead of using a generator when the model has two inputs.

  #20518 closed Dec 25, 2024

• keras==3.7.0 - Attention(... return_attention_scores=True) does not work

  #20680 closed Dec 24, 2024

• Consider changing the import order of `pydot` and its forks

  #20681 closed Dec 23, 2024

6 Issues opened by 6 people

• ValueError: creating distributed tf.Variable with aggregation=MEAN and a non-floating dtype is not supported, please use a different aggregation or dtype

  #20700 opened Dec 28, 2024

• The default setting for aggregation='mean' for variables and optimizers is incorrect.

  #20699 opened Dec 28, 2024

• The default setting for aggregation='mean' in Variable and Optimizer is incorrect.

  #20698 opened Dec 28, 2024

• Variable and Optimizer aggregation='mean' default is incorrect

  #20695 opened Dec 27, 2024

• Loading a model created and trained in Keras 2. in new Keras 3.7 env for inference

  #20687 opened Dec 25, 2024

• Error after upgrade from 3.6 to 3.7

  #20686 opened Dec 24, 2024

16 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.

• Fixing batch_dim_name attribute

  #20674 commented on Dec 27, 2024 • 3 new comments

• [Possible Bug]: Device management issue when using Torch Modules inside Keras Models

  #18411 commented on Dec 23, 2024 • 0 new comments

• 🚀 Contributing to Keras 🚀

  #18442 commented on Dec 23, 2024 • 0 new comments

• EfficientNetLiteB0Backbone in KerasHub

  #20593 commented on Dec 24, 2024 • 0 new comments

• I want to propose a PR for an new ops, which could be in the form of a tritionor a CUDA kerne？

  #20658 commented on Dec 24, 2024 • 0 new comments

• cannot load fine-tuned model

  #19518 commented on Dec 24, 2024 • 0 new comments

• tf.keras compatibility: convert multi-label-classification to Keras Core

  #18438 commented on Dec 24, 2024 • 0 new comments

• Unable to load finetuned keras model.

  #19808 commented on Dec 24, 2024 • 0 new comments

• Untracked embedding layer inside Sequential keras 3 model

  #20676 commented on Dec 26, 2024 • 0 new comments

• TokenAndPositionEmbedding layer warning---Respecting padding mask or destroying it?

  #19794 commented on Dec 26, 2024 • 0 new comments

• [BUG] Conflicting loss_weights implementation in Keras3 for single output case.

  #19780 commented on Dec 27, 2024 • 0 new comments

• Think about Keras3 preference, should I develop my distributed training CTR model based on Jax or TensorFlow?

  #19866 commented on Dec 27, 2024 • 0 new comments

• Keras deserialization Object Error - from keras.saving.legacy.serialization import deserialize_keras_object ImportError: cannot import name 'deserialize_keras_object' from

  #19426 commented on Dec 27, 2024 • 0 new comments

• Unless specified otherwise, keras.Input defaults to 'float32', regardless of the current 'floatx' setting

  #19332 commented on Dec 27, 2024 • 0 new comments

• How to implement train_step with multiple gradient calculations with a JAX backend? Ex. GAN

  #18881 commented on Dec 27, 2024 • 0 new comments

• Request for developer guide: multi-node TPU distributed training with JAX

  #20356 commented on Dec 28, 2024 • 0 new comments