diff --git a/vllm/v1/worker/gpu/attn_utils.py b/vllm/v1/worker/gpu/attn_utils.py index 4f9e5e99a9..5c07860b3b 100644 --- a/vllm/v1/worker/gpu/attn_utils.py +++ b/vllm/v1/worker/gpu/attn_utils.py @@ -238,7 +238,7 @@ def _reshape_attention_kv_cache( page_stride = kv_cache_spec.page_size_bytes // dtype_size num_blocks_dim = inv_order[0] - strides = list(torch.empty(permuted_kv_cache_shape).stride()) + strides = list(torch.empty(permuted_kv_cache_shape, device="meta").stride()) strides[num_blocks_dim] = page_stride kv_cache = torch.as_strided(