Issues

• tf.image.resize_with_pad pads a row/column of zeros due to float32 rounding (e.g. 512x512 -> 224x224)

  type:bug
  tensorflow/tensorflow#129783 · Filter by author bodapatisaikrishna opened on Oct 31, 2026
  

• tf.signal.fftshift and ifftshift fail with TypeError for unknown rank inputs

  type:bug
  tensorflow/tensorflow#129781 · Filter by author Hrushikesh-ramilla opened on Oct 31, 2026
  

• tf.sparse.concat fails with TypeError when sp_inputs is a tuple, and crashes on empty inputs

  type:bug
  tensorflow/tensorflow#129760 · Filter by author Hrushikesh-ramilla opened on Oct 30, 2026
  

• [Google Play Services - TFLite] GPU delegate unavailable on Android 17: tflite_gpu_dynamite no longer enabled (worked on Android 16, same Play services build)

  comp:litetype:support
  tensorflow/tensorflow#129748 · Filter by author alifesoftware-work opened on Oct 30, 2026
  1

  

• tf.signal.hann_window/hamming_window ignore periodic=True for odd window lengths

  2.21.0awaiting PR mergecomp:signaltype:bug
  tensorflow/tensorflow#129743 · Filter by author bodapatisaikrishna opened on Oct 30, 2026
  1

  

• tf.linalg.matrix_rank and lu_* cast float64/complex NumPy inputs to float32 (wrong rank, imaginary part dropped)

  2.21.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129739 · Filter by author bodapatisaikrishna opened on Oct 30, 2026
  1

  

• tf.image.resize gradients fail for half/bfloat16/float64 with antialias or lanczos/gaussian/mitchellcubic, and are silently None for bicubic half/bfloat16

  2.21.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129737 · Filter by author bodapatisaikrishna opened on Oct 30, 2026
  

• tf.image.convert_image_dtype overflows with bfloat16/half and wide ints (bfloat16 1.0 -> uint8 gives 0)

  2.21.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129731 · Filter by author bodapatisaikrishna opened on Oct 30, 2026
  

• MLIR Graph Optimizer fails with 'operand #0 does not dominate this use' on trivial Keras model (TF 2.21.0, macOS arm64)

  2.21.0comp:kerastype:bug
  tensorflow/tensorflow#129711 · Filter by author heinerbilch opened on Oct 29, 2026
  3

  

• tf.image.adjust_saturation turns red pixels gray or black on CPU when blue is a few ulps above green

  2.21.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129605 · Filter by author usmanmateen opened on Oct 27, 2026
  1

  

• [CUDA] illegal memory read on MaxPoolV2 with qint8 NCHW_VECT_C

  2.20.0awaiting PR mergecomp:gpucomp:opstype:bug
  tensorflow/tensorflow#129587 · Filter by author kokol16 opened on Oct 27, 2026
  

• [CUDA] illegal memory read on ResourceScatterAdd with float16

  2.20.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129577 · Filter by author kokol16 opened on Oct 26, 2026
  1

  

• [CUDA] illegal memory read on MaxPoolWithArgmax

  2.20.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129576 · Filter by author kokol16 opened on Oct 26, 2026
  1

  

• [CUDA] illegal memory write on DenseBincount when size is close to INT32_MAX

  2.20.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129575 · Filter by author kokol16 opened on Oct 26, 2026
  

• tf.gather_nd with int16 indices returns uninitialized memory or aborts when indices hold 32,768 or more index tuples

  2.21.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129561 · Filter by author joszamama opened on Oct 26, 2026
  1

  

• TF 2.21.0 source build fails: LLVM checksum mismatch "Checksum was 00b1077e… but wanted 3f986184…"

  2.21.0awaiting PR mergesubtype: ubuntu/linuxtype:build/install
  tensorflow/tensorflow#129543 · Filter by author KEERTHANA-G-R opened on Oct 26, 2026
  1

  

• float32 tf.matmul on rank-3 tensors returns wrong, nondeterministic results on CPU (oneDNN) for batch 2-16 and contraction >= 20000

  2.21.0awaiting PR mergecomp:coretype:bug
  tensorflow/tensorflow#129498 · Filter by author italo-goncalves opened on Oct 23, 2026
  3

  

• overlap_and_add returns non-empty tensor for 0-frame inputs when frame_length > frame_step

  comp:signalstat:awaiting responseTF 2.19type:bug
  tensorflow/tensorflow#129484 · Filter by author Abhirup0 opened on Oct 23, 2026
  2

  

• tf.get_static_value() never folds StridedSlice, so tf.shape(x)[0] has no static value in graph mode

  2.21.0awaiting PR mergecomp:coretype:bug
  tensorflow/tensorflow#129438 · Filter by author Phoebus-Liu opened on Oct 22, 2026
  1

  

• tf.raw_ops.For aborts on GPU for a valid float32 loop while the CPU execution succeeds

  2.21.0comp:gpucomp:opsstat:awaiting tensorflowertype:bug
  tensorflow/tensorflow#129396 · Filter by author lv20031001 opened on Oct 21, 2026
  1

  

• tf.raw_ops.SparseMatrixAdd with complex128 on GPU causes a segmentation fault

  2.21.0comp:opsstat:awaiting tensorflowertype:bug
  tensorflow/tensorflow#129392 · Filter by author lv20031001 opened on Oct 21, 2026
  1

  

• ImportEvent aborts the process for a multi-element event tensor instead of returning an error

  2.21.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129384 · Filter by author lv20031001 opened on Oct 21, 2026
  1

  

• ResourceAccumulatorApplyGradient aborts on accumulator/gradient dtype mismatch

  2.21.0awaiting PR mergecomp:opstype:bug
  tensorflow/tensorflow#129358 · Filter by author lv20031001 opened on Oct 20, 2026
  2

  

• RealDiv accepts integer dtypes in its OpDef, but eager/graph have no integer kernel and reject, while XLA (jit_compile=True) silently does truncating integer division — a dtype-support / semantic-consistency gap

  2.21.0comp:xlastat:awaiting tensorflowertype:bug
  tensorflow/tensorflow#129335 · Filter by author tinywisdom opened on Oct 20, 2026
  2

  

• [Numerical accuracy] Conv2DBackpropInput float32 deviation under severe cancellation on CPU

  comp:opsTF 2.16type:bug
  tensorflow/tensorflow#129166 · Filter by author L4AIsS opened on Oct 16, 2026
  1