learning/annotated_deep_learning_paper_implementations
1
0
Fork 0
mirror of https://github.com/labmlai/annotated_deep_learning_paper_implementations.git synced 2025-08-14 09:31:42 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #226 from MrYxJ/patch-1

Browse Source 
Fix a typo in the formula of ALiBi.
This commit is contained in:
Varuna Jayasiri
2023-11-17 17:42:17 +00:00
committed by GitHub
parent 07b5782a48 830161b299
commit 36a374ed76
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
labml_nn/transformers/alibi/__init__.py
View File

@ -19,7 +19,7 @@ Here's the attention formula for $i$-th token,
\begin{align}
\mathbf{a}_i
&= \text{softmax} \bigg( \mathbf{q}_i \mathbf{K}^\top + m \cdot \big[-(i-1), \dots, 1, 0 \big] \bigg) \\
&= \text{softmax} \bigg( \mathbf{q}_i \mathbf{K}^\top + m \cdot \big[-(i-1), \dots, -1, 0 \big] \bigg) \\
&= \text{softmax} \bigg( \mathbf{q}_i \mathbf{K}^\top + m \cdot \big[0, 1, \dots, (i - 1) \big] \bigg)
\end{align}