[API Compatibility No.361] Param alias for lerp using decorator -part #77170
Conversation
|
你的PR提交成功，感谢你对开源项目的贡献! |
Codecov Report❌ Patch coverage is
❌ Your patch status has failed because the patch coverage (66.66%) is below the target coverage (90.00%). You can increase the patch coverage or adjust the target coverage. Additional details and impacted files@@ Coverage Diff @@
## develop #77170 +/- ##
==========================================
Coverage ? 66.66%
==========================================
Files ? 1
Lines ? 6
Branches ? 0
==========================================
Hits ? 4
Misses ? 2
Partials ? 0 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
/re-run all-failed |
python/paddle/tensor/math.py
Outdated
| if in_dynamic_or_pir_mode(): | ||
| return _C_ops.lerp(x, y, weight) | ||
| if out is None: | ||
| return _C_ops.lerp(x, y, weight) |
python/paddle/tensor/math.py
Outdated
| if in_dynamic_or_pir_mode(): | ||
| return _C_ops.lerp(x, y, weight) | ||
| if out is None: | ||
| return _C_ops.lerp(x, y, weight) |
There was a problem hiding this comment.
这个看起来是可以直接采用下沉的方式。再看看吧。
Manfredss
force-pushed
the
ApiEnhance361
branch
from
8b99dde to
6a4886a
Compare
| weight = paddle.full(shape=[], fill_value=weight, dtype=x.dtype) | ||
|
|
||
| if in_dynamic_or_pir_mode(): | ||
| return _C_ops.lerp(x, y, weight) |
There was a problem hiding this comment.
如果out非None，这里没有处理。这里其实不需要处理，因为_C_ops里面自己会处理是否为None的情况
| if in_dynamic_or_pir_mode(): | ||
| return _C_ops.lerp(x, y, weight) | ||
| if out is None: | ||
| return _C_ops.lerp(x, y, weight, out=out) |
There was a problem hiding this comment.
这个可以看下ops.yaml里，使用 Scalar(double) weight 的形式能否下沉
There was a problem hiding this comment.
没太理解，可以稍微解释一下吗
下沉的主要问题是这个weight可以传float或Tensor，paddle内部内型Scalar可以接收Tensor以及任意的float/int/double等scalar类型，所以可以将op参数改成Scalar类型。这个你可以试一下，参考下allclose就是用scalar实现的，但是还需要改kernel的参数类型，不排除有兼容性问题。如果有问题就还是用装饰器。
4 participants
PR Category
User Experience
PR Types
New features
Description
Param alias for paddle.lerp using decorator
Add out param