[Hardware][ROCM] using current_platform.is_rocm (#9642)

Signed-off-by: wangshuai09 <391746016@qq.com>

vllm/model_executor/models/llama.py

@@ -50,8 +50,8 @@ from vllm.model_executor.model_loader.weight_utils import (
     default_weight_loader, kv_cache_scales_loader, maybe_remap_kv_scale_name)
 from vllm.model_executor.pooling_metadata import PoolingMetadata
 from vllm.model_executor.sampling_metadata import SamplingMetadata
+from vllm.platforms import current_platform
 from vllm.sequence import IntermediateTensors, PoolerOutput
-from vllm.utils import is_hip
 
 from .interfaces import SupportsLoRA, SupportsPP
 from .utils import (AutoWeightsLoader, PPMissingLayer, is_pp_missing_parameter,
@@ -423,7 +423,7 @@ class LlamaModel(nn.Module):
             if not isinstance(self.layers[layer_idx], nn.Identity):
                 layer_self_attn = self.layers[layer_idx].self_attn
 
-            if is_hip():
+            if current_platform.is_rocm():
                 # The scaling factor convention we are assuming is
                 # quantized_value * scaling_factor ~= true_value
                 # which is consistent with the practice of setting