Issues · keras-team/keras · GitHub

🗓️ SIG Keras Community Meetings: Schedule, Invites, and Notes
#22410 · laxmareddyp opened on Mar 12, 2026
🚀 Contributing to Keras 🚀
#18442 · fchollet opened on Jul 14, 2023
52
🗺️ Keras Development Roadmap
#19519 · fchollet opened on Apr 15, 2024
15

Labels Milestones

[Bug] GPU Memory leak in torch backend: `torchtree_impl.py` closure captures tensor containers

#23110

· DavidAriOstenfeldt opened

on Jun 18, 2026

'Operation' object has no attribute 'rematerialized_call' when using RematScope

#23106

· buildwithsuhana opened

on Jun 17, 2026

[Bug] sklearn wrapper warm_start=True does not reuse model instance weights

#23097

· ChiragSW opened

on Jun 15, 2026

[Bug] Segmentation mask visualization docs use one-based class labels but implementation uses zero-based indices

#23088

· ChiragSW opened

on Jun 13, 2026

[Bug] TF tracing fails for distributed training with stateful optimizers

stat:contributions welcome

#23085

· peinguim opened

on Jun 13, 2026

[Bug] `model.evaluate()` reports incorrect metric values in graph mode when dataset has fixed batch dimension in `element_spec`

#23084

· stinodego opened

on Jun 12, 2026

[Bug] `model.fit()` hangs indefinitely after calling `StringLookup.adapt()` on macOS Apple Silicon (CPU) with TF 2.21.0 + Python 3.13

stat:awaiting response from contributor

#23083

· Koji-Minami opened

on Jun 12, 2026

[Bug] TorchDataLoaderAdapter corrupts dict and single-tensor batches on non-Torch backends

#23080

· ChiragSW opened

on Jun 12, 2026

`ops.normalize(order=2)` yields NaN gradients for zero vectors on the torch backend — the L2 fast path bypasses the function's own epsilon guard

#23075

· nmandilaras opened

on Jun 11, 2026

[Bug] `tf.data` class weights fail for omitted classes

#23071

· ChiragSW opened

on Jun 10, 2026

[Bug] Audio resampling truncates before resampling

#23068

· ChiragSW opened

on Jun 10, 2026

Question: Validate Huntr security reports before Huntr's OSS sunset

#23055

· LinZiyuu opened

on Jun 8, 2026