[CI] Fix LM Eval Large Models (H100) (#32423)

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com> (cherry picked from commit bcf2333cd6)
2026-01-15 19:52:49 -05:00
parent c2a37a3cf8
commit 7f42dc20bb
1 changed files with 3 additions and 1 deletions
--- a/vllm/model_executor/layers/quantization/input_quant_fp8.py
+++ b/vllm/model_executor/layers/quantization/input_quant_fp8.py
@@ -98,7 +98,9 @@ class QuantFP8(CustomOp):
            num_token_padding=self.num_token_padding,
            scale_ub=scale_ub,
            use_per_token_if_dynamic=self.use_per_token_if_dynamic,
-            group_shape=self.group_shape if self.static else None,
+            group_shape=(self.group_shape.row, self.group_shape.col)
+            if self.static
+            else None,
        )

    def forward_hip(