Issues

Issues

• Complex gradients of multiply_no_nan, xlogy, xlog1py, xdivy, cumprod, segment_prod and unsorted_segment_prod are conjugated wrongly

  tensorflow/tensorflow#130245 · opened on Nov 12, 2026
  2

• [XLA] A dynamic dimension of run-time size 1 is not broadcast: silent wrong shape

  tensorflow/tensorflow#130232 · opened on Nov 12, 2026

• tf.image.adjust_hue and tf.image.adjust_saturation have no gradient: LookupError under tf.function, silent None in eager

  tensorflow/tensorflow#130207 · opened on Nov 11, 2026
  1

• tf.image.hsv_to_rgb has no gradient: registered NotDifferentiable, tf.GradientTape returns None

  tensorflow/tensorflow#130205 · opened on Nov 11, 2026
  1

• tf.linalg.norm rejects NumPy integer axes (np.int64), while Python ints and np.linalg.norm work

  tensorflow/tensorflow#130195 · opened on Nov 11, 2026

• [docs] tf.nn.avg_pool2d / max_pool2d / max_pool_with_argmax: a length-4 ksize/strides with a non-1 batch entry is rejected ("Pooling is not yet supported on the batch dimension"), which the docs do not mention

  tensorflow/tensorflow#130189 · opened on Nov 11, 2026

• tf.nn.sampled_softmax_loss and tf.nn.nce_loss fail for float16, bfloat16 and float64 inputs

  tensorflow/tensorflow#130178 · opened on Nov 10, 2026

• GPU XLA produces nonfinite Conv1D and LayerNorm weights after one finite-loss training step on Blackwell

  tensorflow/tensorflow#130169 · opened on Nov 10, 2026
  1

• CPU Conv2D segfaults after Eigen temporary allocation failure

  tensorflow/tensorflow#130165 · opened on Nov 10, 2026

• [RFC] Add codespell spelling presubmit for changed Python/Markdown files

  tensorflow/tensorflow#130085 · opened on Nov 8, 2026

• tf.math.polyval fails for Python int coefficients with a float x

  tensorflow/tensorflow#130023 · opened on Nov 6, 2026

• tf.Tensor.__dlpack__ rejects a matching dl_device instead of using the zero-copy path

  tensorflow/tensorflow#129993 · opened on Nov 5, 2026
  3

• Feature Request: Standalone Training Loop Support for Android on CPU (NPU + GPGPU solutions are discussed too)

  tensorflow/tensorflow#129927 · opened on Nov 3, 2026
  13

• [Sparse] Support SparseTensorValue in tf.sparse.bincount and validate_sparse_weights

  tensorflow/tensorflow#129921 · opened on Nov 3, 2026

• [Sparse] Validate input and convert SparseTensorValue in tf.sparse.reduce_sum and tf.sparse.reduce_max

  tensorflow/tensorflow#129916 · opened on Nov 3, 2026

• tf.sparse.softmax fails on non-SparseTensor inputs and does not convert SparseTensorValue

  tensorflow/tensorflow#129908 · opened on Nov 3, 2026
  1

• tf.sparse.maximum and tf.sparse.minimum fail on non-SparseTensor inputs and do not convert SparseTensorValue

  tensorflow/tensorflow#129904 · opened on Nov 3, 2026
  1

• tf.sparse.expand_dims fails on SparseTensorValue and non-SparseTensor inputs

  tensorflow/tensorflow#129901 · opened on Nov 3, 2026
  1

• tf.math.igammac / tf.math.igamma: docs list bfloat16 and half, but there is no kernel for them on CPU or GPU

  tensorflow/tensorflow#129878 · opened on Nov 3, 2026

• tf.image.draw_bounding_boxes raises TypeError when colors argument is omitted in TF 2

  tensorflow/tensorflow#129799 · opened on Nov 1, 2026
  1

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

  tensorflow/tensorflow#129748 · opened on Oct 30, 2026
  1

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

  tensorflow/tensorflow#129605 · opened on Oct 27, 2026
  2

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

  tensorflow/tensorflow#129587 · opened on Oct 27, 2026
  2

• [CUDA] illegal memory read on ResourceScatterAdd with float16

  tensorflow/tensorflow#129577 · opened on Oct 26, 2026
  2

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

  tensorflow/tensorflow#129543 · opened on Oct 26, 2026
  1