[XPU][Tests] Enable test_per_token_group_quant_int8 on XPU (#55681)
Signed-off-by: pmanczak <pawel.manczak@intel.com>
This commit is contained in:
@@ -413,9 +413,12 @@ def test_per_token_group_quant_fp8_packed_large_mn():
|
|||||||
|
|
||||||
@pytest.mark.parametrize("shape", [(32, 128), (64, 256), (16, 512)])
|
@pytest.mark.parametrize("shape", [(32, 128), (64, 256), (16, 512)])
|
||||||
@pytest.mark.parametrize("group_size", [64, 128])
|
@pytest.mark.parametrize("group_size", [64, 128])
|
||||||
@pytest.mark.skipif(not torch.cuda.is_available(), reason="CUDA not available")
|
@pytest.mark.skipif(
|
||||||
|
not (current_platform.is_cuda_alike() or current_platform.is_xpu()),
|
||||||
|
reason="Only test on CUDA/ROCm/XPU.",
|
||||||
|
)
|
||||||
def test_per_token_group_quant_int8(shape, group_size: int):
|
def test_per_token_group_quant_int8(shape, group_size: int):
|
||||||
device = "cuda"
|
device = current_platform.device_type
|
||||||
|
|
||||||
torch.manual_seed(42)
|
torch.manual_seed(42)
|
||||||
num_tokens, hidden_dim = shape
|
num_tokens, hidden_dim = shape
|
||||||
|
|||||||
Reference in New Issue
Block a user