Unify isViewable, handle n-dimensional empty tensors. #8883
Conversation
1) Unifies the two isViewable functions in ATen and TH. 2) Handle n-dimensional empty tensors in the implementation 3) Clarify some comments. This requires an extra copy in the TH case, but that will go away.
gchanan
requested review from
apaszke,
colesbury,
ezyang,
soumith and
zdevito
as code owners
|
Can we unify the THC version somehow too? https://github.com/pytorch/pytorch/blob/master/aten/src/THC/generic/THCTensor.cpp#L242 |
|
why is there a THC version :(. |
|
Yeah sorry I forgot about that one earlier. :( We basically duplicated everything for TH & THC, so... |
|
@pytorchbot retest this please. |
|
We should probably unify THTensor/THCTensor doubletime, but there seem to be some yaks in the way... |
aten/src/TH/THTensor.cpp
Outdated
| // On a high level, | ||
| // 1. separate tensor->size into chunks of dimensions, where the dimensions are | ||
| // ``contiguous'' in each chunk, i.e., stride[i] = size[i+1] * stride[i+1] | ||
| // 2. view_size must be able to be separated into same number of chunks as tensor->size was separated into, |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
| // 2. view_size must be able to be separated into same number of chunks as tensor->size was separated into, | ||
| // where each chunk of view_size has matching ``numel'', i.e., number of subspaces, | ||
| // as the corresponding chunk of tensor->size. | ||
| at::optional<std::vector<int64_t>> |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
ezyang
left a comment
ezyang
left a comment
There was a problem hiding this comment.
Always hard to tell what has changed when you move code around and make changes at the same time :)
|
@pytorchbot retest this please. |
* upstream/master: (42 commits) [c10d] No default device for ProcessGroupGloo (pytorch#8888) Fix default values for affine= in the docstrings of InstanceNormXd (pytorch#8895) Stop making dynamic allocations of PinnedMemoryAllocator. (pytorch#8896) [C++ API] Rework optimization package (pytorch#8815) Mention MPICH_MAX_THREAD_SAFETY=multiple. (pytorch#8580) Unify isViewable, handle n-dimensional empty tensors. (pytorch#8883) Add pos_weight argument to nn.BCEWithLogitsLoss (pytorch#5660) (pytorch#6856) [build] Enable clang-specific warnings only when using clang (pytorch#8869) Fix cmake cudnn autodetection (pytorch#8891) [c10d] Fix link order for building C++ tests (pytorch#8889) directly add_subdirectory(nanopb) from torch CMakeLists (pytorch#8870) [C++ API] Bag of fixes (pytorch#8843) [build] Raise in cmake when seeing NVCC{9/9.1} + GCC6 combo (pytorch#8863) Create avg_pool1d in ATen (pytorch#8880) throw error when grid_sample is passed unsupported mode (pytorch#8884) Allow autograd to work even when the shape of values cannot be determined (pytorch#8641) Make at::Tensor::to() const (pytorch#8839) [auto] Update onnx to 458c521 - Fix typo (onnx/onnx#1143) onnx/onnx@458c521 [Caffe2] Fix gradient_check on in-place ops (pytorch#8828) Fix as_strided_backward (pytorch#8721) ...
3 participants
This requires an extra copy in the TH case, but that will go away.