Skip to content

【Hackathon 5th No.17】 为 Paddle 新增 pdist API（update） #760

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
luotao1 merged 8 commits into from
Nov 30, 2023
Merged

【Hackathon 5th No.17】 为 Paddle 新增 pdist API（update） #760

Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
749f6fe
fix rfc
cocoshe Oct 24, 2023
833e511
add other possible strategy
cocoshe Nov 2, 2023
ebd008a
modify api param: A to x
cocoshe Nov 20, 2023
7cf4d4e
Merge branch 'master' into householderRFC
cocoshe Nov 20, 2023
2c50a85
fix
cocoshe Nov 22, 2023
019ff1e
Merge branch 'householderRFC' of https://github.com/cocoshe/community…
cocoshe Nov 22, 2023
5faf569
Merge branch 'master' of https://github.com/PaddlePaddle/community in…
cocoshe Nov 24, 2023
f4b8160
update pdist api param
cocoshe Nov 24, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 1 addition & 5 deletions rfcs/APIs/20230926_api_design_for_pdist.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -366,16 +366,12 @@ Scipy利用现有API组合实现，PyTorch则在底层重写cpp算子。

API的设计为:

`paddle.pdist(x, p=2.0, compute_mode="use_mm_for_euclid_dist_if_necessary", name=None)`
`paddle.pdist(x, p=2.0, name=None)`

Args：

+ x(Tensor): 严格为 shape=[M, N] 的 Tensor
+ p(float, 可选): 为p-范数对应的p值，默认为2.0
+ compute_mode(str, 可选): 默认为`use_mm_for_euclid_dist_if_necessary`（组合已有API过程中用到了`paddle.cdist`，当`p=2.0`时，可以设置`compute_mode`利用矩阵运算进行优化）
+ `compute_mode=use_mm_for_euclid_dist_if_necessary`时，当p=2.0且M>25时使用矩阵乘法计算距离
+ `compute_mode=use_mm_for_euclid_dist`时，当p=2.0时使用矩阵乘法计算距离
+ `compute_mode=donot_use_mm_for_euclid_dist`时，不使用矩阵乘法计算距离
+ name(str, 可选): 操作的名称(默认为None)

Return：
Expand Down