Small fixes for slice edge cases #476
Merged
Conversation
This file contains hidden or 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
Contributor
|
Please add previously-failing tests which now pass to demonstrate the new functionality :) |
Collaborator
Author
|
Unfortunately this case goes from xfail to xfail, as it results in 0 size tensors which still crash the refbackend #448 Do we want to create tests for xfail on Optional type coverage? |
Contributor
|
is there a way to hack around that temporarily? for example, to cat the zero size tensor with a tensor of size 1 so the final tensor is of size 1? |
dan-garvey
force-pushed
the
slice_fixes
branch
from
e20f920 to
093a996
Compare
Collaborator
Author
|
Have they tried cloning you yet |
silvasean
approved these changes
Dec 14, 2021
Contributor
silvasean
left a comment
silvasean
left a comment
There was a problem hiding this comment.
Go ahead and merge after addressing nit :)
| ]) | ||
| def forward(self, x): | ||
| return x[5:5, 3:3, -1:] | ||
| #TODO: remove hacky cat tensor once refbackend supports 0 size dim |
dan-garvey
force-pushed
the
slice_fixes
branch
from
093a996 to
7ae06ea
Compare
qedawkins
pushed a commit
to nod-ai/torch-mlir
that referenced
this pull request
Oct 3, 2022
* Add krnl.load and knrl.store Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Lower krnl load/store to affine load/store or std load/store Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Edit optimize-memory-pools pass to adapt with the new changes Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Use krnl load/store in onnx-to-krnl pass Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Edit lit tests Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Revise createKrnlLoad/Store Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Add lit tests for krnl.load/store lowering Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Edit lit tests for loop op Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Remaining affine.load in IndexExpr Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Mark std.load/store and affine.load/store illegal at the krnl level Signed-off-by: Tung D. Le <tung@jp.ibm.com> * Use krnl.load/store in the LRN lowering Signed-off-by: Tung D. Le <tung@jp.ibm.com> Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Adds a check for optional type, and covers the case where end < start. Thanks to @cathyzhyi for pointing these out.