-
Notifications
You must be signed in to change notification settings - Fork 89
Open
Labels
contribution welcomeWe welcome code contributions for thisWe welcome code contributions for thisgood first issueGood for newcomersGood for newcomersmodule: torchlibRelated to the torch/aten function lib in developmentRelated to the torch/aten function lib in development
Description
When converting a graph with repeat_interleave, I'm hitting the following error
<class 'TypeError'>: aten_repeat_interleave_self_int() got an unexpected keyword argument 'output_size'
Looks like onnxscript is missing support for the argument?
onnxscript/onnxscript/function_libs/torch_lib/ops/core.py
Lines 7512 to 7515 in 70e751a
| def aten_repeat_interleave_self_int( | |
| self: TensorType, repeats: int, dim: Optional[int] = None | |
| ) -> TensorType: | |
| """repeat_interleave.self_int(Tensor self, SymInt repeats, int? dim=None, *, SymInt? output_size=None) -> Tensor |
justinchuby and wenbingl
Metadata
Metadata
Assignees
Labels
contribution welcomeWe welcome code contributions for thisWe welcome code contributions for thisgood first issueGood for newcomersGood for newcomersmodule: torchlibRelated to the torch/aten function lib in developmentRelated to the torch/aten function lib in development