Skip to content

Initialize logical_element_type before sharding in XLATensor::Data #4418

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
yeounoh merged 1 commit into from
Jan 5, 2023
Merged

Initialize logical_element_type before sharding in XLATensor::Data #4418

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions torch_xla/csrc/tensor.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -107,13 +107,13 @@ class XLATensor : public torch::lazy::LazyTensor {
~Data();

std::shared_ptr<View> view;
// TODO: remove this in favor of torch::lazy::Shape within ir_value.
c10::optional<at::ScalarType> logical_element_type;
// The user provided sharding spec is attached to `XLATensor::Data`
// and all sharding look-up should refer to it as source of truth.
// A copy of the sharding spec is attached to the IR node via
// `SetShardingSpec` and also during the sync tensor collection.
ShardingSpecPtr sharding;
// TODO: remove this in favor of torch::lazy::Shape within ir_value.
c10::optional<at::ScalarType> logical_element_type;
};

static XLATensorPtr Create(const at::Tensor& tensor,
Expand Down