Repository navigation

[pull] master from tensorflow:master - #8855

Merged

pull[bot] merged 30 commits into

Cache-Cloud:masterfrom

tensorflow:master

Sep 28, 2026

Merged

[pull] master from tensorflow:master#8855
pull[bot] merged 30 commits into
Cache-Cloud:masterfrom
tensorflow:master

Commits on Sep 7, 2026
- Add AArch64 oneDNN async module repo
  cfRod
  committed
  87372d7
Commits on Sep 22, 2026
- Validate axis bounds in experimental.numpy reduction functions
  kaivalya-cyber
  committed
  b599ff6
- Handle scalar and 0-d array axes in reduction axis validation
  kaivalya-cyber
  committed
  bb4d7e2
- Expect ValueError for out-of-bounds reduction axis in testReductionOfOutOfBoundsAxis
  kaivalya-cyber
  committed
  f137c3e
- Reject negative integer powers on GPU
  GodlyDonuts
  committed
  a712eb4
- Simplify GPU Pow validation callback
  GodlyDonuts
  committed
  094f260
- Retain GPU validation tensor safely
  GodlyDonuts
  committed
  6a1d236
Commits on Sep 23, 2026
- Use valid integer exponents in generated GPU Pow tests
  GodlyDonuts
  committed
  05f42d2
Commits on Sep 24, 2026
- Declare Abseil status dependency for GPU integer Pow
  GodlyDonuts
  committed
  8084b92
Commits on Sep 25, 2026
Commits on Sep 28, 2026
- [XLA] Upgrade nanobind to da42c8a5890eff09e672e7cfd953627655aa3e28
  
  hawkinsp
  authored and
  tensorflower-gardener
  committed
  6ad8816
- [XLA:GPU][AMD] Disable topk_special_floats_test to unblock presubmits.
  
  pifon2a
  authored and
  tensorflower-gardener
  committed
  1a01b64
- [XLA:CPU] Remove redundant passes from the new XTile compilation pipeline
  
  pifon2a
  authored and
  tensorflower-gardener
  committed
  659a0a5
- [XLA] Extend HLO kDot with block scaling and sparsity.
  
  sashabu
  authored and
  tensorflower-gardener
  committed
  5ad2263
- [XLA:CPU] Cap vector tile size and unroll large or batched vector.transpose ops
  
  pifon2a
  authored and
  tensorflower-gardener
  committed
  7d5b984
- Reverts f87634eaff1171e67c8d98cb04198cf168fc4c98
  tensorflower-gardener
  committed
  98b4ae0
- [XLA] Cache xla.range for apply_indexing op.
  
  pifon2a
  authored and
  tensorflower-gardener
  committed
  66585ca
- Optimize tflite::reference_ops::StridedSlice via contiguous dimension coalescing and WriteN.
  tensorflower-gardener
  committed
  3df6868
- Merge pull request #127873 from GodlyDonuts:codex/reject-negative-gpu-integer-pow
  tensorflower-gardener
  committed
  1bed4d3
- PR #49586: [ROCm]Add rocm_platform_test.
  
  linchen1-robot
  authored and
  tensorflower-gardener
  committed
  89ba87d
- [XLA:GPU] Fix out-of-bounds index handling in ScatterDeterminismExpander
  
  akuegel
  authored and
  tensorflower-gardener
  committed
  9725c54
- Merge pull request #127606 from kaivalya-cyber:fix/np-reduce-axis-bounds
  tensorflower-gardener
  committed
  01f1f55
- Mutex guard cuDNN compilation handle.
  
  derdrdirk
  authored and
  tensorflower-gardener
  committed
  e1a22a9
- Add constant folding pass before fusing convolutions.
  
  derdrdirk
  authored and
  tensorflower-gardener
  committed
  96aebc2
- [XLA:GPU] Allow a symbolic TilingSpace to be cloned into another MLIRContext
  
  Moerafaat
  authored and
  tensorflower-gardener
  committed
  57e6215
- Merge pull request #128029 from bhushanasati25:fix-nested-forward-accumulator
  tensorflower-gardener
  committed
  749b510
- Merge pull request #127109 from cfRod:onednn-aarch64-async-propagation
  tensorflower-gardener
  committed
  bf313a7

Navigation Menu

Repository navigation

[pull] master from tensorflow:master - #8855

[pull] master from tensorflow:master#8855
pull[bot] merged 30 commits into
Cache-Cloud:masterfrom
tensorflow:master

Commits on Sep 7, 2026

Add AArch64 oneDNN async module repo

Commits on Sep 22, 2026

Validate axis bounds in experimental.numpy reduction functions

Handle scalar and 0-d array axes in reduction axis validation

Expect ValueError for out-of-bounds reduction axis in testReductionOfOutOfBoundsAxis

Reject negative integer powers on GPU

Simplify GPU Pow validation callback

Retain GPU validation tensor safely

Commits on Sep 23, 2026

Use valid integer exponents in generated GPU Pow tests

Commits on Sep 24, 2026

Declare Abseil status dependency for GPU integer Pow

Commits on Sep 25, 2026

Account for input aliasing in forward-mode function caches

Merge remote-tracking branch 'origin/master' into fix/np-reduce-axis-bounds

Guard GPU-only integer Pow header in CPU builds

Iterate forward-mode cache inputs without concatenation

Commits on Sep 28, 2026

[XLA] Upgrade nanobind to da42c8a5890eff09e672e7cfd953627655aa3e28

[XLA:GPU][AMD] Disable topk_special_floats_test to unblock presubmits.

[XLA:CPU] Remove redundant passes from the new XTile compilation pipeline

[XLA] Extend HLO kDot with block scaling and sparsity.

[XLA:CPU] Cap vector tile size and unroll large or batched vector.transpose ops

Reverts f87634eaff1171e67c8d98cb04198cf168fc4c98

[XLA] Cache xla.range for apply_indexing op.

Optimize tflite::reference_ops::StridedSlice via contiguous dimension coalescing and WriteN.

Merge pull request #127873 from GodlyDonuts:codex/reject-negative-gpu-integer-pow

PR #49586: [ROCm]Add rocm_platform_test.

[XLA:GPU] Fix out-of-bounds index handling in ScatterDeterminismExpander

Merge pull request #127606 from kaivalya-cyber:fix/np-reduce-axis-bounds

Mutex guard cuDNN compilation handle.

Add constant folding pass before fusing convolutions.

[XLA:GPU] Allow a symbolic TilingSpace to be cloned into another MLIRContext

Merge pull request #128029 from bhushanasati25:fix-nested-forward-accumulator

Merge pull request #127109 from cfRod:onednn-aarch64-async-propagation