AddOp(upsample_bicubic2d) | feat(torchlib) #1208
Conversation
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #1208 +/- ##
==========================================
+ Coverage 78.72% 78.79% +0.06%
==========================================
Files 118 118
Lines 15602 15645 +43
Branches 2468 2476 +8
==========================================
+ Hits 12283 12327 +44
Misses 2911 2911
+ Partials 408 407 -1 ☔ View full report in Codecov by Sentry. |
Test Results 24 files ± 0 24 suites ±0 1h 31m 58s ⏱️ - 5m 6s For more details on these failures, see this check. Results for commit 8222049. ± Comparison against base commit 1fa1ed6. This pull request removes 21 and adds 26 tests. Note that renamed tests count towards both.♻️ This comment has been updated with latest results. |
justinchuby
changed the title
| """upsample_bicubic2d.vec(Tensor input, SymInt[]? output_size, bool align_corners, float[]? scale_factors) -> Tensor""" | ||
| """upsample_bicubic2d(Tensor self, SymInt[2] output_size, bool align_corners, float? scales_h=None, float? scales_w=None) -> Tensor""" # pylint: disable=pointless-string-statement |
There was a problem hiding this comment.
Should we split them into two functions?
There was a problem hiding this comment.
Onnx.Resize() only accept either output_size or scales, it is error if both provided.
So for the second signature, when the output_size have been provided, the scales_h and scales_w will be ignored
There was a problem hiding this comment.
Do we handle when scales_h and scales_w are not None for upsample_bicubic2d?
There was a problem hiding this comment.
Didn't handled.
Because from this signature upsample_bicubic2d(Tensor self, SymInt[2] output_size, bool align_corners, float? scales_h=None, float? scales_w=None) -> Tensor, the output_size must be provided, then we will ignore scales_h and scales_w.
No description provided.