Skip to content

Fix reference tensor copy for PR2673 #3

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
ptrendx merged 1 commit into from
Mar 12, 2026
Merged

Fix reference tensor copy for PR2673 #3

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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: 2 additions & 2 deletions tests/pytorch/test_fusible_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -178,7 +178,7 @@ def make_reference_and_test_tensors(
test = test.dequantize()

# Make sure reference and test tensors match each other
ref.copy_(test)
ref.copy_(test.to(dtype=ref.dtype))

ref.requires_grad_(requires_grad)
test.requires_grad_(requires_grad)
Expand Down Expand Up @@ -1956,7 +1956,7 @@ def test_dropout(
)
with torch.no_grad():
x_test += 1
x_ref.copy_(x_test)
x_ref.copy_(x_test.to(dtype=x_ref.dtype))
dy_ref, dy_test = make_reference_and_test_tensors(
shape,
test_dtype=dtype,
Expand Down
2 changes: 1 addition & 1 deletion tests/pytorch/test_quantized_tensor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -173,7 +173,7 @@ def make_reference_and_test_tensors(
raise ValueError(f"Unsupported quantization scheme ({quantization})")

# Make sure reference and test tensors match each other
ref.copy_(test)
ref.copy_(test.to(dtype=ref.dtype))

ref.requires_grad_(requires_grad)
test.requires_grad_(requires_grad)
Expand Down
2 changes: 1 addition & 1 deletion transformer_engine/pytorch/quantized_tensor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -552,7 +552,7 @@ def __torch_dispatch__(cls, func, types, args, kwargs=None):
dst.quantize_(src)
else:
if isinstance(src, QuantizedTensor):
dtype = src.dtype
dtype = dst.dtype
if dtype not in (torch.float32, torch.float16, torch.bfloat16):
dtype = torch.float32
src = src.dequantize(dtype=dtype)
Expand Down