This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
added ravel/unravel operators #11025
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
|
Could you make sure the behavior is exactly the same with numpy for the same configuration? |
|
Changed the behavior to match numpy: A single multi index is given by a column of the matrix (not by a row). @piiswrong is that what you recommended in your comments? |
anirudh2290
reviewed
May 23, 2018
| DType *unravelled, DType *ravelled) { | ||
| index_t idx(ravelled[i]); | ||
| #pragma unroll | ||
| for (int j = ndim; j--; ) { |
There was a problem hiding this comment.
I find this style a little confusing and would prefer for(int j = ndim; j >= 0; j--) . What is the motivation ?
There was a problem hiding this comment.
Just a style issue. I have seen people using either variant.
asmushetzel
force-pushed
the
ravel2
branch
2 times, most recently
from
08485cf to
749d8e8
Compare
|
Is this ready to merge? We have an important use case for these ops and would like to test them soon. |
|
Ready from my point of view. |
piiswrong
suggested changes
May 29, 2018
src/operator/tensor/ravel.cc
Outdated
| .set_attr<FResourceRequest>("FResourceRequest", [](const NodeAttrs& attrs) | ||
| { return std::vector<ResourceRequest>{ResourceRequest::kTempSpace}; }) | ||
| .set_attr<nnvm::FListInputNames>("FListInputNames", [](const NodeAttrs& attrs) | ||
| { return std::vector<std::string>{"A"}; } ) |
src/operator/tensor/ravel.cc
Outdated
| .set_attr<FResourceRequest>("FResourceRequest", [](const NodeAttrs& attrs) | ||
| { return std::vector<ResourceRequest>{ResourceRequest::kTempSpace}; }) | ||
| .set_attr<nnvm::FListInputNames>("FListInputNames", [](const NodeAttrs& attrs) | ||
| { return std::vector<std::string>{"A"}; } ) |
src/operator/tensor/ravel.cc
Outdated
| .set_attr<nnvm::FInferType>("FInferType", ElemwiseType<1, 1>) | ||
| .set_attr<FCompute>("FCompute<cpu>", RavelForward<cpu>) | ||
| .set_attr<nnvm::FGradient>("FGradient", MakeZeroGradNodes) | ||
| .add_argument("A", "NDArray-or-Symbol", "Batch of multi-indices") |
src/operator/tensor/ravel.cc
Outdated
| .set_attr<nnvm::FInferType>("FInferType", ElemwiseType<1, 1>) | ||
| .set_attr<FCompute>("FCompute<cpu>", UnravelForward<cpu>) | ||
| .set_attr<nnvm::FGradient>("FGradient", MakeZeroGradNodes) | ||
| .add_argument("A", "NDArray-or-Symbol", "Array of flat indices") |
asmushetzel
force-pushed
the
ravel2
branch
3 times, most recently
from
fe3a77c to
f9b9b13
Compare
|
Changed input names from "A" to "data" as requested. |
rahul003
pushed a commit
to rahul003/mxnet
that referenced
this pull request
Jun 4, 2018
zheng-da
pushed a commit
to zheng-da/incubator-mxnet
that referenced
this pull request
Jun 28, 2018
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Description
This resolves #10203 by implementing ravel_multi_index and unravel_index operators. The operators perform the same functionality as the corresponding operators in numpy.
This enables writing fully symbolic operations that involve ravel/unravel of indices.
Checklist
Essentials
Please feel free to remove inapplicable items for your PR.