diff --git a/vllm/executor/executor_base.py b/vllm/executor/executor_base.py
index fb76276bb4b34..242690f8e1b8f 100644
--- a/vllm/executor/executor_base.py
+++ b/vllm/executor/executor_base.py
@@ -8,11 +8,11 @@ from typing import (Any, Awaitable, Callable, Dict, List, Optional, Set, Tuple,
 import torch.nn as nn
 from typing_extensions import TypeVar
 
+import vllm.platforms
 from vllm.config import VllmConfig
 from vllm.logger import init_logger
 from vllm.lora.request import LoRARequest
 from vllm.model_executor.layers.sampler import SamplerOutput
-from vllm.platforms import current_platform
 from vllm.prompt_adapter.request import PromptAdapterRequest
 from vllm.sequence import ExecuteModelRequest, PoolerOutput
 from vllm.utils import make_async
@@ -108,8 +108,8 @@ class ExecutorBase(ABC):
         """
         # NOTE: This is logged in the executor because there can be >1 workers.
         logger.info("# %s blocks: %d, # CPU blocks: %d",
-                    current_platform.dispatch_key, num_gpu_blocks,
-                    num_cpu_blocks)
+                    vllm.platforms.current_platform.dispatch_key,
+                    num_gpu_blocks, num_cpu_blocks)
         max_concurrency = (num_gpu_blocks * self.cache_config.block_size /
                            self.model_config.max_model_len)
         logger.info("Maximum concurrency for %s tokens per request: %.2fx",
diff --git a/vllm/executor/ray_utils.py b/vllm/executor/ray_utils.py
index 7b30155971a6d..33c0a25803ca6 100644
--- a/vllm/executor/ray_utils.py
+++ b/vllm/executor/ray_utils.py
@@ -7,10 +7,10 @@ from typing import TYPE_CHECKING, Dict, List, Optional, Tuple, Union
 
 import msgspec
 
+import vllm.platforms
 from vllm.config import ParallelConfig
 from vllm.executor.msgspec_utils import decode_hook, encode_hook
 from vllm.logger import init_logger
-from vllm.platforms import current_platform
 from vllm.sequence import ExecuteModelRequest, IntermediateTensors
 from vllm.utils import get_ip
 from vllm.worker.worker_base import WorkerWrapperBase
@@ -54,10 +54,10 @@ try:
 
         def get_node_and_gpu_ids(self) -> Tuple[str, List[int]]:
             node_id = ray.get_runtime_context().get_node_id()
-            device_key = current_platform.ray_device_key
+            device_key = vllm.platforms.current_platform.ray_device_key
             if not device_key:
                 raise RuntimeError("current platform %s does not support ray.",
-                                   current_platform.device_name)
+                                   vllm.platforms.current_platform.device_name)
             gpu_ids = ray.get_runtime_context().get_accelerator_ids(
             )[device_key]
             return node_id, gpu_ids
diff --git a/vllm/platforms/cuda.py b/vllm/platforms/cuda.py
index 991d55ac861a4..9deb0294668ec 100644
--- a/vllm/platforms/cuda.py
+++ b/vllm/platforms/cuda.py
@@ -334,10 +334,10 @@ class NvmlCudaPlatform(CudaPlatformBase):
             if (len(set(device_names)) > 1
                     and os.environ.get("CUDA_DEVICE_ORDER") != "PCI_BUS_ID"):
                 logger.warning(
-                    "Detected different devices in the system: \n%s\nPlease"
+                    "Detected different devices in the system: %s. Please"
                     " make sure to set `CUDA_DEVICE_ORDER=PCI_BUS_ID` to "
                     "avoid unexpected behavior.",
-                    "\n".join(device_names),
+                    ", ".join(device_names),
                 )