Skip to content

Commit 8eb8cd1

Browse files
authored
Name Correction op_full_like (#12755)
### Summary Operator and dtype selective build requires operator names to match operators in use. This changes corrects the name for `full_like` operator. ### Test plan OpenAI's whisper-tiny encoder model requires the full_like operator. Modified the `examples/selective_build/test_selective_build.sh` script to use the pte from this model. Model previous crashed with missing operator, but now completes without error. I.e. ran `bash examples/selective_build/test_selective_build.sh cmake`.
1 parent a30bbdd commit 8eb8cd1

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

kernels/portable/cpu/op_full_like.cpp

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -50,7 +50,7 @@ Tensor& full_like_out(
5050

5151
ScalarType out_type = out.scalar_type();
5252

53-
constexpr auto name = "scalar_tensor.out";
53+
constexpr auto name = "full_like.out";
5454

5555
ET_SWITCH_REALHBBF16_TYPES(out_type, ctx, name, CTYPE_OUT, [&] {
5656
auto opt_val_casted =

0 commit comments

Comments
 (0)