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 13, 12:00 UTC
[BugFix] Fix RuntimeError in PixtralHFAttention on CPU/XPU (#28444)
Signed-off-by: Lin, Fanli <fanli.lin@intel.com>
Fanli Lin committed
Nov 11, 2025 at 23:29 UTC
b886068056a05857f796909d2f8573b36fc668a5
1 file changed
+1
-1
vllm/model_executor/models/pixtral.py
+1
-1
@@ -1109,7 +1109,7 @@ class PixtralHFAttention(nn.Module):
1109
)
1110
out = out.transpose(1, 2)
1111
1112
- out = out.view(batch, patches, self.n_heads * self.head_dim)
1112
+ out = out.reshape(batch, patches, self.n_heads * self.head_dim)
1113
attn_output, _ = self.o_proj(out)
1114
1115
return attn_output, None