Feat lazy tensor indexing #9334
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
wyg1997
requested review from
hjchen2,
BBuf,
daquexian and
jackalcooper
as code owners
hjchen2
reviewed
Nov 7, 2022
wyg1997
commented
Nov 15, 2022
| Scalar value_scalar = functional::PyUnpackScalar(value); | ||
| value_tensor = ASSERT_PTR( | ||
| functional::Constant(Shape({}), value_scalar, tensor->dtype(), ASSERT(tensor->device()))); | ||
| std::shared_ptr<Tensor> value_tensor; |
Comment on lines
+566
to
+567
| // NOTE: LocalToGlobal should be called in eager mode | ||
| LazyMode::Guard lazy_mode_disabled_guard(/*is_enabled*/ false); |
There was a problem hiding this comment.
LocalToGlobal 只能在 eager 模式下调用
Comment on lines
106
to
112
| /*static*/ Maybe<void> SliceUpdateOp::InferPhysicalTensorDesc(user_op::InferContext* ctx) { | ||
| return InferLogicalTensorDesc(ctx); | ||
| const user_op::TensorDesc& ref_desc = ctx->InputTensorDesc("ref", 0); | ||
| auto* y_desc = ctx->MutOutputTensorDesc("y", 0); | ||
| y_desc->set_shape(ref_desc.shape()); | ||
| y_desc->set_is_dynamic(ref_desc.is_dynamic()); | ||
| return Maybe<void>::Ok(); | ||
| } |
There was a problem hiding this comment.
之前 SliceUpdate 的物理 Tensor 推导是错误的，它支持 S + B -> S，是不能和逻辑 shape 推导共用推导函数（逻辑推导函数中有一些 shape 的检察，物理 tensor shape 推导不需要）
hjchen2
approved these changes
Nov 21, 2022
Speed stats: |
|
View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9334/ |
Speed stats: |
|
View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9334/ |
wyg1997
requested review from
oneflow-ci-bot
and removed request for
oneflow-ci-bot
Speed stats: |
marigoold
approved these changes
Nov 23, 2022
Speed stats: |
|
View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/9334/ |
farmerzhang1
pushed a commit
that referenced
this pull request
Nov 30, 2022
* feat(boxing): collective_boxing slice_boxing support 0size tensor * test(Indexing): add lazy tensor basic indexing * add MaskTensor judgement * format code * feat(TensorIndexing): support lazy advance getitem indexing * feat(Indexing): support lazy indexing for lazy_tensor and free_tensor * fix(Indexing): fix indexing test bug * test(Indexing): test all advance indexing * test(GlobalIndexing): fix eager global indexing bug * test(Indexing): support combined indexing * add last test cases * fix merge bug * fix lazy mode guard * test(Indexing): refine set scalar value test * test(Indexing): enable all bool tensor index setitem * decrease test time * refine 0size shape judgement * add comment
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
lazy tensor indexing 支持。
主要处理了：