Skip to content

torch.repeat cast short ints to int32 #390

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ev-br merged 2 commits into from
Feb 17, 2026
Merged

torch.repeat cast short ints to int32 #390

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
955d146
BUG: torch: cast short ints in repeat
ev-br Feb 17, 2026
c5a0028
TST: unskip test_repeat on torch and numpy
ev-br Feb 17, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions array_api_compat/torch/_aliases.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -573,6 +573,10 @@ def count_nonzero(

# "repeat" is torch.repeat_interleave; also the dim argument
def repeat(x: Array, repeats: int | Array, /, *, axis: int | None = None) -> Array:
if isinstance(repeats, torch.Tensor) and repeats.dtype in (torch.int8, torch.int16):
# torch rejects short integers for the `repeat` argument:
# https://github.com/pytorch/pytorch/issues/151311
repeats = repeats.to(torch.int32)
return torch.repeat_interleave(x, repeats, axis)


Expand Down
2 changes: 0 additions & 2 deletions numpy-1-22-xfails.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -143,8 +143,6 @@ array_api_tests/test_operators_and_elementwise_functions.py::test_where_with_sca
array_api_tests/test_has_names.py::test_has_names[creation-from_dlpack]
array_api_tests/test_signatures.py::test_func_signature[from_dlpack]
array_api_tests/test_signatures.py::test_array_method_signature[__dlpack__]
# uint64 repeats not supported
array_api_tests/test_manipulation_functions.py::test_repeat

# 2024.12 support
array_api_tests/test_signatures.py::test_func_signature[bitwise_and]
Expand Down
2 changes: 0 additions & 2 deletions numpy-1-26-xfails.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,8 +39,6 @@ array_api_tests/meta/test_hypothesis_helpers.py::test_symmetric_matrices
# 2023.12 support
array_api_tests/test_signatures.py::test_func_signature[from_dlpack]
array_api_tests/test_signatures.py::test_array_method_signature[__dlpack__]
# uint64 repeats not supported
array_api_tests/test_manipulation_functions.py::test_repeat

# 2024.12 support
array_api_tests/test_data_type_functions.py::TestResultType::test_with_scalars
Expand Down
4 changes: 0 additions & 4 deletions numpy-dev-xfails.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,10 +8,6 @@ array_api_tests/test_data_type_functions.py::test_finfo[complex64]
array_api_tests/test_signatures.py::test_func_signature[vecdot]
array_api_tests/test_signatures.py::test_extension_func_signature[linalg.vecdot]

# 2023.12 support
# uint64 repeats not supported
array_api_tests/test_manipulation_functions.py::test_repeat

# Stubs have a comment: (**note**: libraries may return ``NaN`` to match Python behavior.); Apparently, NumPy does just that
array_api_tests/test_special_cases.py::test_binary[floor_divide(x1_i is +infinity and isfinite(x2_i) and x2_i > 0) -> +infinity]
array_api_tests/test_special_cases.py::test_binary[floor_divide(x1_i is +infinity and isfinite(x2_i) and x2_i < 0) -> -infinity]
Expand Down
4 changes: 0 additions & 4 deletions numpy-xfails.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,10 +8,6 @@ array_api_tests/test_data_type_functions.py::test_finfo[complex64]
array_api_tests/test_signatures.py::test_func_signature[vecdot]
array_api_tests/test_signatures.py::test_extension_func_signature[linalg.vecdot]

# 2023.12 support
# uint64 repeats not supported
array_api_tests/test_manipulation_functions.py::test_repeat

# Stubs have a comment: (**note**: libraries may return ``NaN`` to match Python behavior.); Apparently, NumPy does just that
array_api_tests/test_special_cases.py::test_binary[floor_divide(x1_i is +infinity and isfinite(x2_i) and x2_i > 0) -> +infinity]
array_api_tests/test_special_cases.py::test_binary[floor_divide(x1_i is +infinity and isfinite(x2_i) and x2_i < 0) -> -infinity]
Expand Down
2 changes: 0 additions & 2 deletions torch-xfails.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -139,8 +139,6 @@ array_api_tests/test_data_type_functions.py::test_finfo_dtype
array_api_tests/test_data_type_functions.py::test_iinfo_dtype

# 2023.12 support
# https://github.com/pytorch/pytorch/issues/151311: torch.repeat_interleave rejects short integers
array_api_tests/test_manipulation_functions.py::test_repeat
# Argument 'device' missing from signature
array_api_tests/test_signatures.py::test_func_signature[from_dlpack]
# Argument 'max_version' missing from signature
Expand Down
Loading