[pull] master from tensorflow:master - #8800

Merged

pull[bot] merged 54 commits into

Cache-Cloud:masterfrom

tensorflow:master

Sep 2, 2026

Commits on Mar 18, 2026

Validate scalar shape for input_min/input_max in QuantizeAndDequantizeV3
ayushozha
committed
fea2878

Commits on Apr 7, 2026

Fix OOB write in MaxPoolGradWithArgmax (CWE-787)
adithyan-ak
committed
7003dde

Commits on May 19, 2026

Enforce runtime vector rank invariants in StringNGrams
jmestwa-coder
committed
9ebce0d

Commits on May 29, 2026

Merge branch 'master' into fix/maxpool-grad-argmax-oob-write
adithyan-ak
authored
61ea1fa

Commits on Jul 2, 2026

Add regression test for MaxPoolGradWithArgmax OOB argmax indices
adithyan-ak
committed
7b7afc0

Commits on Jul 5, 2026

test: accept graph quantize scalar shape error
ayushozha
committed
145da4b

Commits on Jul 9, 2026

handle graph-mode shape inference error in StringNGrams rank test
jmestwa-coder
committed
832f9e1

Commits on Jul 17, 2026

reject empty row_splits when decoding RaggedTensorVariant
madib06ops
committed
32772f9

Commits on Jul 20, 2026

Commits on Jul 23, 2026

Validate batch_index rank in Unbatch before accessing its dimensions
vishwakt
committed
f22a66e

Commits on Jul 24, 2026

test(scatter_nd): address review feedback on rank-validation tests
Yigtwxx
committed
448305d

Commits on Aug 13, 2026

Validate reverse axes for scalar and empty inputs
vishwakt
committed
efadb4b

Commits on Aug 15, 2026

fix(rnn): bounds check seq_len_max in BlockLSTM ops to prevent OOB access
veerjain-1
committed
86d7965

Commits on Aug 23, 2026

test: add bounds validation tests for BlockLSTM seq_len_max
veerjain-1
committed
c841dac
Merge master and resolve conflicts
veerjain-1
committed
592bd0d
fix(test): wrap long lines in rnn_cell_test.py to pass PyLint
veerjain-1
committed
f8fbb96

Commits on Aug 24, 2026

fix(test): update expected error message for seq_len_max bounds check
veerjain-1
committed
40ad276

Commits on Aug 26, 2026

Validate data rank in Unbatch and index rank in UnbatchGrad
vishwakt
committed
67fa3b0

Commits on Aug 29, 2026

Fix integer overflow and type mapping in Unicode operations
opabhijeet
committed
116b81d
Add missing Tsplits=int32 tests for UnicodeDecodeWithOffsets
opabhijeet
committed
30b8178
Trigger CI
opabhijeet
committed
5a6c469
Add value assertions for char_values and char_to_byte_starts
opabhijeet
committed
815924d
Fix pylint line-too-long errors
opabhijeet
committed
aeb4a60
upgrade offsets to int64 and fix python formatting
opabhijeet
committed
1edb19c

Commits on Sep 1, 2026

Fix tensorflow/lite/delegates/ynnpack/attention_bench build in open-source.

vksnk
authored and
tensorflower-gardener
committed
b7f54ac
Reverts 7b34ce2de44d6b34e35ae10d2f5dbd7218e173a9

akuegel
authored and
tensorflower-gardener
committed
cf5232b
Fix memory leak in GradientTape::ComputeGradient on backward function error.
tensorflower-gardener
committed
4c493e0
Add support for feature group convolutions (`feature_group_count > 1`) and batch group convolutions (`batch_group_count > 1`) in Mosaic TPU and JAX Pallas.

blakehechtman
authored and
tensorflower-gardener
committed
d31f980
PR #47176: [ROCM] Wire all MORI collectives through a dedicated CollectivesFacade

pemeliya
authored and
tensorflower-gardener
committed
8eee663
Strip bazel-out prefix from generated #include directives.
tensorflower-gardener
committed
98d6125
[XLA:MSA] Clamp a view use's prefetch deadline to the view's own time

patrick-toulme
authored and
tensorflower-gardener
committed
e3ab2a2
[IFRT] Add RemapPlan constructor with input_devices_for_output_map and no mappings

junwhanahn
authored and
tensorflower-gardener
committed
2b01994
This is an automatic update to a device compatibility allowlist.

apaiu-dev
authored and
tensorflower-gardener
committed
822a287
Automated Code Change
tensorflower-gardener
committed
fc34c82
Support tuple entry parameters in MakeDataflowConstrainedArguments.

bhatuzdaname
authored and
tensorflower-gardener
committed
22d32c4
PR #47769: Reject transpose tiles exceeding shared memory during Triton tile selection

mfrancepillois
authored and
tensorflower-gardener
committed
5853a3b
Allow using higher alignment for dmas for dynamic shape metadata.

pschuh
authored and
tensorflower-gardener
committed
be9bcee
[XLA] Add xla_deduplicate_backend_configs_min_size flag to control backend config deduplication

thcmbs
authored and
tensorflower-gardener
committed
1e6ebb9
Remove the no-op flow of uploading and using autotune results.

shyamliagrawal
authored and
tensorflower-gardener
committed
aa0530d
Replace `TF_ASSERT_OK_AND_ASSIGN` with `ASSERT_OK_AND_ASSIGN` in `pjrt_c_api_client_test.cc`.

zacmustin
authored and
tensorflower-gardener
committed
f0f71dc
Integrate LLVM at llvm/llvm-project@cbc5a226cbf8
tensorflower-gardener
committed
3cf2b22
Merge pull request #125400 from veerjain-1:fix-blocklstmgrad-oob
tensorflower-gardener
committed
786097d
Prefactor for supporting remote tile SPMEM transfers.

emilyfertig
authored and
tensorflower-gardener
committed
2458545
Merge pull request #123379 from Yigtwxx:fix/scatter-nd-rank-validation-93680
tensorflower-gardener
committed
4604684
Merge pull request #125269 from vishwakt:fix-reverse-scalar-axis-validation
tensorflower-gardener
committed
6862b7b
Merge pull request #123865 from vishwakt:fix-unbatch-index-rank-crash
tensorflower-gardener
committed
b86b8e5
Merge pull request #117891 from opabhijeet:fix-unicode-splits-dtype
tensorflower-gardener
committed
f7eabcb
Merge pull request #123446 from madib06ops:ragged-variant-empty-splits
tensorflower-gardener
committed
28cb2ac
Merge pull request #109355 from ayushozha:fix/quantize-dequantize-v3-input-validation
tensorflower-gardener
committed
d05dcb5
Merge pull request #118868 from jmestwa-coder:string-ngrams-runtime-rank-validation
tensorflower-gardener
committed
fa336cc
[SDY][Bug-Fix] Fix Shardy shard_map export for nested ManualComputationOp across calls

Varcho
authored and
tensorflower-gardener
committed
e22f2b2

Commits on Sep 2, 2026

Merge pull request #115326 from adithyan-ak:fix/maxpool-grad-argmax-oob-write
tensorflower-gardener
committed
7be031e
Propagate min/max reduction identity element constraints in ConstraintPropagator.

bhatuzdaname
authored and
tensorflower-gardener
committed
9c6fd26