Skip to content

[xpu][test] Port 2 test/quantization_{qat, quant_api} UT files to intel XPU #3351

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
merged 4 commits into from
Nov 18, 2025
Merged

[xpu][test] Port 2 test/quantization_{qat, quant_api} UT files to intel XPU #3351

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
9c1de99
port 2 files to intel XPU
zxd1997066 Nov 18, 2025
c5cdd7a
port 2 files to intel XPU
zxd1997066 Nov 18, 2025
5b76453
update
zxd1997066 Nov 18, 2025
079d4b3
update
zxd1997066 Nov 18, 2025
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
10 changes: 6 additions & 4 deletions test/quantization/test_qat.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,12 +98,14 @@
)
from torchao.utils import (
_is_fbgemm_gpu_genai_available,
get_current_accelerator_device,
is_fbcode,
is_sm_at_least_89,
)

# TODO: put this in a common test utils file
_CUDA_IS_AVAILABLE = torch.cuda.is_available()
_DEVICE = get_current_accelerator_device()
Copy link
Collaborator

@liangan1 liangan1 Nov 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Directly use "_DEVICE is None" in the following changes? No need to define _GPU_IS_AVAILABLE.

Copy link
Contributor Author

@zxd1997066 zxd1997066 Nov 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done



class Sub(torch.nn.Module):
Expand Down Expand Up @@ -347,7 +349,7 @@ def _set_ptq_weight(
group_size,
)
q_weight = torch.ops.aten._convert_weight_to_int4pack(
q_weight.to("cuda"),
q_weight.to(_DEVICE),
qat_linear.inner_k_tiles,
)
ptq_linear.weight = q_weight
Expand Down Expand Up @@ -600,13 +602,13 @@ def _assert_close_4w(self, val, ref):
print(mean_err)
self.assertTrue(mean_err < 0.05)

@unittest.skipIf(not _CUDA_IS_AVAILABLE, "skipping when cuda is not available")
@unittest.skipIf(_DEVICE is None, "skipping when gpu is not available")
def test_qat_4w_primitives(self):
n_bit = 4
group_size = 32
inner_k_tiles = 8
scales_precision = torch.bfloat16
device = torch.device("cuda")
device = torch.device(_DEVICE)
dtype = torch.bfloat16
torch.manual_seed(self.SEED)
x = torch.randn(100, 256, dtype=dtype, device=device)
Expand Down Expand Up @@ -699,7 +701,7 @@ def test_qat_4w_quantizer(self):

group_size = 32
inner_k_tiles = 8
device = torch.device("cuda")
device = torch.device(_DEVICE)
dtype = torch.bfloat16
torch.manual_seed(self.SEED)
m = M().to(device).to(dtype)
Expand Down
Loading
Loading