Repository navigation

[pull] master from tensorflow:master - #8861

Merged

pull[bot] merged 70 commits into

Cache-Cloud:masterfrom

tensorflow:master

Sep 30, 2026

Merged

[pull] master from tensorflow:master#8861
pull[bot] merged 70 commits into
Cache-Cloud:masterfrom
tensorflow:master

Commits on Jul 23, 2026
- Reject malformed empty boxes and box_index in crop-and-resize kernels
  vishwakt
  committed
  a55459e
- Add separator before shape in boxes and box_index rank error messages
  vishwakt
  committed
  b7d372e
Commits on Aug 23, 2026
- Add missing strict dep for image_grad_test_base
  vishwakt
  committed
  4e66734
Commits on Aug 26, 2026
- Reject SplitV size_splits overflow instead of aborting
  vishwakt
  committed
  77637cc
- Skip the SplitV overflow test when XLA is enabled
  vishwakt
  committed
  86e8344
- Use disable_xla decorator for the SplitV overflow test
  vishwakt
  committed
  b47c91e
Commits on Aug 29, 2026
- Merge master into fix-splitv-size-overflow
  vishwakt
  committed
  be5e051
Commits on Sep 11, 2026
- Merge master into fix-splitv-size-overflow
  vishwakt
  committed
  14f3ab7
Commits on Sep 21, 2026
- Stop exporting statically linked LLVM and MLIR symbols from libtensorflow_framework
  vishwakt
  committed
  2ca8a3a
- Match LLVM and MLIR by namespace instead of by substring
  vishwakt
  committed
  f2c41de
Commits on Sep 22, 2026
Commits on Sep 23, 2026
- Validate filter element count against 32-bit limit in GPU conv transformations
  RohithPariki
  committed
  779544a
Commits on Sep 24, 2026
- Address PR feedback: Move bounds check before allocation, use direct comparison, and remove memory-heavy python test
  RohithPariki
  committed
  b21b554
Commits on Sep 25, 2026
Commits on Sep 26, 2026
- Drop the empty-input early return in ParseAndCheckBoxSizes
  vishwakt
  committed
  3122d28
- Format the crop-and-resize changes
  vishwakt
  committed
  668a864
Commits on Sep 27, 2026
- Validate axes in experimental.numpy transpose
  kaivalya-cyber
  committed
  7113623
Commits on Sep 28, 2026
Commits on Sep 29, 2026
Commits on Sep 30, 2026
- Fix NameError in rot90 unsigned-axis tests: use onp.uint32 instead of np.uint32
  kaivalya-cyber
  committed
  8e5c138
- Merge master into fix/np-rot90-axes
  kaivalya-cyber
  committed
  eb512d3
- Use plain abs() instead of builtins.abs in rot90 duplicate check
  kaivalya-cyber
  committed
  398deb9
- Merge master into fix/np-rot90-axes
  kaivalya-cyber
  committed
  f83995b
- Disable libnvjitlink disk caching via the -no-cache option.
  
  beckerhe
  authored and
  tensorflower-gardener
  committed
  6dedc2c
- [IFRT Proxy] Avoid client error masking in client disconnection
  
  hyeontaek
  authored and
  tensorflower-gardener
  committed
  b1e0513
- Merge pull request #127683 from kaivalya-cyber:fix/np-rot90-axes
  tensorflower-gardener
  committed
  3a6e837
- [XLA:GPU] fix number of symbols when simplifying tiles
  
  metaflow
  authored and
  tensorflower-gardener
  committed
  4b89a61
- Automated Code Change
  tensorflower-gardener
  committed
  863f92a
- [XLA:GPU] Skip collective symmetric buffer tests on non-Hopper architectures.
  
  sohaibiftikhar
  authored and
  tensorflower-gardener
  committed
  4e9a743
- [XLA:GPU] Add PerDeviceState container for per-device runtime state.
  
  sohaibiftikhar
  authored and
  tensorflower-gardener
  committed
  29b9c9f
- Merge pull request #127780 from vishwakt:fix-framework-llvm-symbol-visibility
  tensorflower-gardener
  committed
  e1d06e3
- Merge pull request #128227 from MaddipatlaChetan24:patch-43
  tensorflower-gardener
  committed
  9402391
- Merge pull request #128143 from kaivalya-cyber:fix/np-transpose-axes-validation
  tensorflower-gardener
  committed
  9d72633
- Automated Code Change
  tensorflower-gardener
  committed
  40bfb8d
- [XLA:GPU] Run the cuDNN frontend graph warmup on a private stream to avoid race conditions when sharing the stream between multiple threads.
  
  derdrdirk
  authored and
  tensorflower-gardener
  committed
  c3f7240
- PR #49717: Export control_predecessors() to python
  
  dlcompilers-infra-bot
  authored and
  tensorflower-gardener
  committed
  2816dca
- PR #49741: [ROCm] Remove dead ROCm version checks from GemmRewriter
  
  mminutoli
  authored and
  tensorflower-gardener
  committed
  bbdf033
- [XLA:GPU] Add REDUCE_SCATTER to xla_gpu_experimental_use_collective_kernels.
  
  PatriosTheGreat
  authored and
  tensorflower-gardener
  committed
  17bf4fb
- PR #49300: [XLA:GPU] Inline scan computation to_apply calls
  
  akhilgoe
  authored and
  tensorflower-gardener
  committed
  0d920ad
- [XLA:GPU] tighter constraint checks in indexing map
  
  metaflow
  authored and
  tensorflower-gardener
  committed
  e1d6d57
- [NFC] Compute GPU instruction annotation titles and metadata once per instruction.
  
  EusebioDM
  authored and
  tensorflower-gardener
  committed
  25acadd
- Merge pull request #123864 from vishwakt:fix-crop-and-resize-empty-crash
  tensorflower-gardener
  committed
  8ebd54a
- Merge pull request #127968 from RohithPariki:fix/87457-gpu-conv-filter-transform-overflow
  tensorflower-gardener
  committed
  480682d
- Merge pull request #126161 from vishwakt:fix-splitv-size-overflow
  tensorflower-gardener
  committed
  c443ce9
- [XLA:GPU] Support disabled VMM API case.
  
  PatriosTheGreat
  authored and
  tensorflower-gardener
  committed
  9fe191d
- [XLA:GPU]: Move memcpy inside all-gather
  
  sohaibiftikhar
  authored and
  tensorflower-gardener
  committed
  645e3ca
- [StableHLO] Fix windows link error in evalRunParallel
  
  Srishti Srivastava
  authored and
  tensorflower-gardener
  committed
  5a1f0fd
- PR #49719: [tsl:concurrency] Deduplicate AsyncValue type ids by type name across DSOs
  
  ezhulenev
  authored and
  tensorflower-gardener
  committed
  270a245
- PR #47649: [XLA:GPU] Keep scatter window writes coalesced in ScatterSimplifier and the transpose folding
  
  kodlan
  authored and
  tensorflower-gardener
  committed
  b02620e
- PR #49767: [ROCm] Restrict usage of system env variables for rbe builds in bzl
  
  alekstheod
  authored and
  tensorflower-gardener
  committed
  717f958
- Skip NVTX-only annotation work when no profiler domain is attached.
  
  EusebioDM
  authored and
  tensorflower-gardener
  committed
  54a6c3a
- Removing deprecated macros in XLA:GPU E2E tests and autotuner
  
  Moerafaat
  authored and
  tensorflower-gardener
  committed
  4706a67
- Migrate deprecated TF assertion macros in XLA:GPU codegen and Triton
  
  Moerafaat
  authored and
  tensorflower-gardener
  committed
  e1e8f86
- Automated Code Change
  tensorflower-gardener
  committed
  a3165c2

Commits

Commits on Jul 23, 2026

Commits on Aug 23, 2026

Commits on Aug 26, 2026

Commits on Aug 29, 2026

Commits on Sep 11, 2026

Commits on Sep 21, 2026

Commits on Sep 22, 2026

Commits on Sep 23, 2026

Commits on Sep 24, 2026

Commits on Sep 25, 2026

Commits on Sep 26, 2026

Commits on Sep 27, 2026

Commits on Sep 28, 2026

Commits on Sep 29, 2026

Commits on Sep 30, 2026