[BUGFIX][ROCM] ViT FlashAttention on ROCm (no GFX9) and contiguous on qwen3vl ROCm TORCH_SDPA (#27190)

Signed-off-by: JartX <sagformas@epdcenter.es>
Co-authored-by: tjtanaa <tunjian.tan@embeddedllm.com>

vllm/model_executor/models/qwen2_vl.py

@@ -462,6 +462,12 @@ class Qwen2VisionAttention(nn.Module):
             ).contiguous()
         elif self.attn_backend == _Backend.TORCH_SDPA:
             # Execute attention entry by entry for speed & less VRAM.
+            from vllm.platforms import current_platform
+
+            if current_platform.is_rocm():
+                q = q.contiguous()
+                k = k.contiguous()
+                v = v.contiguous()
             outputs = []
             for i in range(1, len(cu_seqlens)):
                 start_idx = cu_seqlens[i - 1]