Review changes requested

Signed-off-by: Flavia Beo <[email protected]>

vllm/config.py

@@ -108,8 +108,6 @@ class ModelConfig:
             can not be gathered from the vllm arguments.
         config_format: The config format which shall be loaded.
             Defaults to 'auto' which defaults to 'hf'.
-        bert_config: tokenizationconfiguration dictionary for a given 
-            Sentence Transformer BERT model.
         mm_processor_kwargs: Arguments to be forwarded to the model's processor
             for multi-modal data, e.g., image processor.
         pooling_type: Used to configure the pooling method in the embedding 

vllm/engine/arg_utils.py

@@ -15,6 +15,7 @@
                          SpeculativeConfig, TaskOption, TokenizerPoolConfig)
 from vllm.executor.executor_base import ExecutorBase
 from vllm.logger import init_logger
+from vllm.model_executor.layers.pooler import PoolingType
 from vllm.model_executor.layers.quantization import QUANTIZATION_METHODS
 from vllm.transformers_utils.config import (
     maybe_register_config_serialize_by_value)
@@ -850,7 +851,7 @@ def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
 
         parser.add_argument(
             '--pooling-type',
-            choices=['LAST', 'ALL', 'CLS', 'STEP', 'MEAN'],
+            choices=[pt.name for pt in PoolingType],
             default=None,
             help='Used to configure the pooling method in the embedding model.'
         )