[fix] engine args import

Signed-off-by: AlpinDale <alpindale@gmail.com>

Author: AlpinDale

aphrodite/modeling/model_loader/tensorizer.py

@@ -27,7 +27,7 @@
 from aphrodite.utils.import_utils import PlaceholderModule
 
 if TYPE_CHECKING:
-    from aphrodite.engine.arg_utils import EngineArgs
+    from aphrodite.engine.args_tools import EngineArgs
 
 try:
     from tensorizer import DecryptionParams, EncryptionParams, TensorDeserializer, TensorSerializer

benchmarks/benchmark_long_document_qa_throughput.py

@@ -34,17 +34,16 @@
                     (Optional, default: 0)
 
 In the meantime, it also supports all the Aphrodite engine args to initialize the 
-LLM engine. You can refer to the `aphrodite.engine.arg_utils.EngineArgs` for more
+LLM engine. You can refer to the `aphrodite.engine.args_tools.EngineArgs` for more
 details.
 """
 
 import dataclasses
 import random
 import time
 
-from aphrodite.engine.arg_utils import EngineArgs
-
 from aphrodite import LLM, SamplingParams
+from aphrodite.engine.args_tools import EngineArgs
 from aphrodite.utils.argparse_utils import FlexibleArgumentParser
 
 

benchmarks/benchmark_prefix_caching.py

@@ -31,10 +31,10 @@
 import random
 import time
 
-from aphrodite.engine.arg_utils import EngineArgs
 from transformers import PreTrainedTokenizerBase
 
 from aphrodite import LLM, SamplingParams
+from aphrodite.engine.args_tools import EngineArgs
 from aphrodite.utils.argparse_utils import FlexibleArgumentParser
 
 try:

benchmarks/benchmark_prioritization.py

@@ -6,9 +6,9 @@
 import random
 import time
 
-from aphrodite.engine.arg_utils import EngineArgs
 from transformers import AutoTokenizer, PreTrainedTokenizerBase
 
+from aphrodite.engine.args_tools import EngineArgs
 from aphrodite.utils.argparse_utils import FlexibleArgumentParser