Update axlearn/common/attention.py

Co-authored-by: Mark Lee <[email protected]>
apple · Jan 15, 2025 · 0be3a83 · 0be3a83
1 parent 7aa6dd5
commit 0be3a83
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/axlearn/common/attention.py b/axlearn/common/attention.py
@@ -1231,7 +1231,7 @@ def forward(
         Args:
             positions: A tensor representing the token position IDs.
                 The shape is [batch_size, seq_len].
-            max_seq_len: Max length of sequence, required if positions is not provided
+            max_seq_len: Max length of sequence, required if positions is not provided.
 
         Returns:
             Rotary Positional Embedding. Shape is [seq_len, dim].