Clean up imports (#5467)

python/sglang/srt/managers/detokenizer_manager.py

@@ -14,7 +14,6 @@
 """DetokenizerManager is a process that detokenizes the token ids."""
 
 import dataclasses
-import json
 import logging
 import os
 import signal

python/sglang/srt/managers/mm_utils.py

@@ -1,7 +1,8 @@
 """
-    Multi-modality utils
+Multi-modality utils
 """
 
+import logging
 from abc import abstractmethod
 from typing import Callable, List, Optional, Tuple
 
@@ -12,11 +13,11 @@ from sglang.srt.managers.schedule_batch import (
     MultimodalDataItem,
     MultimodalInputs,
     global_server_args_dict,
-    logger,
 )
 from sglang.srt.model_executor.forward_batch_info import ForwardBatch
 from sglang.srt.utils import print_warning_once
-from sglang.utils import logger
+
+logger = logging.getLogger(__name__)
 
 
 class MultiModalityDataPaddingPattern:

python/sglang/srt/managers/multimodal_processor.py

@@ -5,8 +5,6 @@ import logging
 import pkgutil
 from functools import lru_cache
 
-from transformers import PROCESSOR_MAPPING
-
 from sglang.srt.managers.multimodal_processors.base_processor import (
     BaseMultimodalProcessor,
 )

python/sglang/srt/managers/multimodal_processors/base_processor.py

@@ -8,8 +8,6 @@ from typing import List, Optional
 
 import numpy as np
 import PIL
-from decord import VideoReader, cpu
-from PIL import Image
 from transformers import BaseImageProcessorFast
 
 from sglang.srt.managers.schedule_batch import Modality
@@ -102,6 +100,9 @@ class BaseMultimodalProcessor(ABC):
         """
         estimate the total frame count from all visual input
         """
+        # Lazy import because decord is not available on some arm platforms.
+        from decord import VideoReader, cpu
+
         # Before processing inputs
         estimated_frames_list = []
         for image in image_data: