rename to need_attn_weights to match elsewhere (pytorch#97102)

Change variable spelling from `need_atten_weights` to `need_attn_weights` to match naming convention elsewhere in pytorch. Pull Request resolved: pytorch#97102 Approved by: https://github.com/drisspg
walterddr · Mar 22, 2023 · 11114ab · 11114ab
1 parent 7a8b691
commit 11114ab
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/aten/src/ATen/native/nested/NestedTensorTransformerFunctions.h b/aten/src/ATen/native/nested/NestedTensorTransformerFunctions.h
@@ -89,15 +89,15 @@ TORCH_API Tensor flash_attention_helper(
     const Tensor& key,
     const Tensor& value,
     double dropout_p,
-    bool need_atten_weights,
+    bool need_attn_weights,
     bool is_causal);
 
 TORCH_API std::tuple<Tensor, Tensor> mem_efficient_helper_nested_unpacked(
     const Tensor& query,
     const Tensor& key,
     const Tensor& value,
     double dropout_p,
-    bool need_atten_weights,
+    bool need_attn_weights,
     bool is_causal);
 } // namespace native
 } // namespace at