A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 2, 04:45 UTC
[XPU] [Bug] remove xpuw4a16 output size check (#44168)
Signed-off-by: Zhu, Zufang <zufang.zhu@intel.com> Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>
zofia committed
Jun 2, 2026 at 22:26 UTC
6314de8bad6f28e235652585d63d75bd38d51dab
1 file changed
-7
vllm/model_executor/kernels/linear/mixed_precision/xpu.py
-7
@@ -51,13 +51,6 @@ class XPUwNa16LinearKernel(MPLinearKernel):
51
"XPUwNa16, supported sizes are multiples of 32",
52
)
53
54
- if c.partition_weight_shape[1] % 32 != 0:
55
- return (
56
- False,
57
- f"Output size ({c.partition_weight_shape[1]}) not supported by "
58
- "XPUWNA16, supported sizes are multiples of 32",
59
- )
60
-
54
return True, None
55
56
def process_weights_after_loading(self, layer: torch.nn.Module):