tensorflow/swift-apis

Issues

• differentiableReduce error

  #941 opened a year ago by jackharmer
  3

• Extrude ShapedArray into a separate package

  #1183 opened 3 years ago by seadiem
  1

• In Colab using 'x10_training_loop' leads to "error: Couldn't lookup symbols:"

  #1016 opened 4 years ago by mikowals
  6

• Requesting assistance

  #1185 opened 3 years ago by philipturner
  36

• Implementation of Metal/Accelerate/DirectX backend

  #1190 opened 3 years ago by philipturner
  0

• Compiling on release mode

  #1189 opened 3 years ago by philipturner
  1

• TensorFlow binary was not compiled to use: SSE4.2 AVX AVX2 FMA

  #935 opened 4 years ago by RahulBhalley
  5

• ShapedArray<String>.description causes crash

  #914 opened 4 years ago by ematejska
  1

• uncaught exception of type Xbyak

  #908 opened 4 years ago by ematejska
  1

• Slow scalar performance even using withDevice(.cpu)

  #903 opened 4 years ago by ematejska
  1

• What's the expected output from Device.allDevices ?

  #1089 opened 4 years ago by xihui-wu
  4

• Make ShapedArray.description's maxScalarCountPerLine user-customizable

  #1168 opened 4 years ago by xanderdunn
  9

• Remove `ParameterlessLayer` conformance workarounds

  #1037 opened 4 years ago by texasmichelle
  1

• Expose all devices.

  #1059 opened 4 years ago by texasmichelle
  1

• Eager Tensors should be placed on the device they are assigned

  #1103 opened 4 years ago by BradLarson
  0

• [summary] Add annotations to `Recurrent`

  #1074 opened 4 years ago by texasmichelle
  1

• Remaining Bidirectional RNN merge modes

  #1098 opened 4 years ago by kongzii
  2

• GPU memory growth and logging level flags aren't being set when X10 is used before eager mode

  #1094 opened 4 years ago by BradLarson
  2

• tf.contrib.layers.flatten(x) alternative function for TensorFlow V0.8.0

  #1105 opened 4 years ago by engrfaiz
  1

• Bidirectional RNN

  #943 opened 4 years ago by kongzii
  3

• Bug when using .concatenated

  #1087 opened 4 years ago by kongzii
  2

• some mutating tensor operations are not @differentiable

  #1048 opened 4 years ago by marcrasi
  7

• [summary] Better formatting

  #1075 opened 4 years ago by texasmichelle
  0

• Lift Tensor.dimensionGathering() out of swift-models and into swift-apis

  #1058 opened 4 years ago by BradLarson
  0

• sigmoidCrossEntropy is not differentiable

  #1038 opened 4 years ago by tanmayb123
  4

• Colab crash using `.differentiableReduce()` on `[Layer]` appended with `+=`

  #1032 opened 4 years ago by mikowals
  1

• GroupNorm crashes on x10

  #1031 opened 4 years ago by s1ddok
  2

• The module file format is too new to be used by this version of the debugger.

  #1041 opened 4 years ago by RahulBhalley
  1

• cmake on MacOS - error: cannot load underlying module for 'XCTest'

  #1028 opened 4 years ago by mikowals
  3

• Add layer wrapper type for `Collection` types

  #1021 opened 4 years ago by dan-zheng
  0

• cmake build fails with bundled libraries - no such module 'x10_xla_tensor_wrapper'

  #1001 opened 4 years ago by mikowals
  0

• device mismatch in pullback of Tensor(_, deviceAndPrecisionLike:)

  #973 opened 4 years ago by mikowals
  4

• On macOS, simple models can trigger a segfault within X10

  #993 opened 4 years ago by BradLarson
  0

• Tensor(linearSpaceFrom:, to:, count:, on:) produces incorrect number of values on X10

  #980 opened 4 years ago by BradLarson
  0

• Remove dependency on TFE_InferShapes and TFE_GetServerDef

  #970 opened 4 years ago by ematejska
  2

• X10 Tensor Performance for LSTMs

  #948 opened 4 years ago by tanmayb123
  4

• X10 optimizer support for `Tensor`

  #979 opened 4 years ago by texasmichelle
  0

• GRUCell internal tensors undersized for inputSize != hiddenSize

  #936 opened 4 years ago by lindyhoptrader
  2

• Test `Slices`

  #968 opened 4 years ago by dabrahams
  0

• Support argsort()

  #964 opened 4 years ago by ddxgz
  7

• How to get target batch variable using TrainingEpochs

  #961 opened 4 years ago by kushukla
  3

• XLA_SAVE_TENSORS_FILE flag slows performance significantly

  #952 opened 4 years ago by mikowals
  1

• Initial `PrintX10Metrics()` call must be early in code execution

  #954 opened 4 years ago by mikowals
  3

• No such module 'TensorFlow'

  #949 opened 4 years ago by RahulBhalley
  3

• Implement shape inference for resource/variant handles.

  #921 opened 4 years ago by ematejska
  1

• Make LazyTensors thread safe.

  #922 opened 4 years ago by ematejska
  2

• Revisit the heuristic to allow materialization during shape inference in lazy tensor

  #918 opened 4 years ago by ematejska
  2

• Debugging and Profiling Tools for Lazy Tensors

  #933 opened 4 years ago by ematejska
  2

• Support complex numbers in matmul.

  #932 opened 4 years ago by ematejska
  0

• Improve TensorFlow runtime errors

  #904 opened 4 years ago by ematejska
  0