Skip to content

[cherry-pick]Add sparse_attention warning (#4083) #4095

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
merged 2 commits into from
Nov 17, 2021
Merged

[cherry-pick]Add sparse_attention warning (#4083) #4095

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3345e9d
Add sparse_attention warning (#4083)
Liu-xiandong Nov 12, 2021
5e49a2a
Merge branch 'release/2.2' into Add_sparse_attention_warnings
Liu-xiandong Nov 16, 2021
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
3 changes: 3 additions & 0 deletions docs/api/paddle/nn/functional/sparse_attention_cn.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,9 @@ sparse_attention

其中，``Q``，``K``，``V`` 表示注意力模块的三个输入参数。这三个参数的维度是一样的。 ``d`` 代表这三个参数的最后一个维度的大小。

.. warning::
目前该API只在CUDA11.3及以上版本中使用。

参数：
:::::::::
- query (Tensor) - 输入的Tensor，代表注意力模块中的 ``query`` ，这是一个4维Tensor，形状为 :[batch_size, num_heads, seq_len, head_dim]，数据类型为float32或float64。
Expand Down