Skip to content

Commit a0b629f

Browse files
committed
add "_gemm_input_role" to dunder slots
Signed-off-by: Masaki Kozuki <[email protected]>
1 parent d4b2f33 commit a0b629f

File tree

1 file changed

+2
-1
lines changed

1 file changed

+2
-1
lines changed

torchao/float8/float8_tensor.py

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -293,8 +293,9 @@ class Float8Tensor(torch.Tensor):
293293
_scale: torch.Tensor
294294
_orig_dtype: torch.dtype
295295
_linear_mm_config: LinearMMConfig
296+
_gemm_input_role: GemmInputRole
296297
_axiswise_dim: Optional[int]
297-
__slots__ = ["_data", "_scale", "_orig_dtype", "_linear_mm_config"]
298+
__slots__ = ["_data", "_scale", "_orig_dtype", "_linear_mm_config", "_gemm_input_role", "_axiswise_dim"]
298299

299300
def __new__(
300301
cls,

0 commit comments

Comments
 (0)