【Hackathon No.7】为 Paddle 新增 apply API -part #59374
Conversation
|
你的PR提交成功，感谢你对开源项目的贡献! |
如果这部分单测测到了，可以补充下本地截图
|
yangguohao
force-pushed
the
apply_tensor
branch
from
4efabdc to
4178851
Compare
yangguohao
force-pushed
the
apply_tensor
branch
from
4178851 to
a4ccbaa
Compare
添加了单测 |
| >>> x = paddle.to_tensor(5.) | ||
| >>> f = lambda x: 3*x+2 | ||
| >>> x.apply_(f) | ||
| >>> print(x) | ||
| Tensor(shape=[], dtype=float32, place=Place(cpu), stop_gradient=True, | ||
| 17.) |
There was a problem hiding this comment.
can the shape of the input data in example code be larger, such as [3, 3]
There was a problem hiding this comment.
are we support other data types? eg float16, bfloat16 and float64, should add more case in example code
There was a problem hiding this comment.
same issue in example code in def apply below
There was a problem hiding this comment.
done add the related example code
|
|
||
| class TestTensorApplyAPI(unittest.TestCase): | ||
| def setUp(self): | ||
| self.x = paddle.to_tensor([1, 2, 3, 4, 5], stop_gradient=True) |
There was a problem hiding this comment.
we need test input tensor in both CPU and GPU
There was a problem hiding this comment.
are we support other data types? eg float16, bfloat16 and float64. if yes, should add more case in here
There was a problem hiding this comment.
Done add the test case
yangguohao
force-pushed
the
apply_tensor
branch
from
32fe7c8 to
fca8389
Compare
|
请提交对应的中文文档 |
我发现好像 CI 里 paddle.seed 固定种子数后输出的随机值和我电脑本地的输出不一样，打印的时候会出现问题。 |
|
@yangguohao 请解决下冲突 |
done 等 CI 完应该就 ok 了 |
luotao1
changed the title
* add tensor apply * fix * fix 2023-11-27 * fix * fix V2 * add apply in Variable * add apply in newir * add test * fix * fix2 * fix example code * change shape * fix docs * fix docs
PR types
New features
PR changes
APIs
Description
add tensor.apply for Paddle
RFC:PaddlePaddle/community#761