[Docs] Fix warnings in mkdocs build (continued) (#24791)

Signed-off-by: Zerohertz <ohg3417@gmail.com>

vllm/model_executor/models/zamba2.py

@@ -909,8 +909,8 @@ class Zamba2ForCausalLM(nn.Module, HasInnerState, IsHybrid):
             prefix: Optional prefix for parameter names
         
         Raises:
-            AssertionError: If prefix caching is enabled
-            (not supported by Mamba)
+            AssertionError: If prefix caching is enabled 
+                (not supported by Mamba)
         """
         config = vllm_config.model_config.hf_config
         cache_config = vllm_config.cache_config

vllm/transformers_utils/config.py

@@ -679,20 +679,21 @@ def get_hf_file_to_dict(file_name: str,
 
 
 @cache
-def get_pooling_config(model: str, revision: Optional[str] = 'main'):
+def get_pooling_config(model: str,
+                       revision: Optional[str] = 'main') -> Optional[dict]:
     """
     This function gets the pooling and normalize
     config from the model - only applies to
     sentence-transformers models.
 
     Args:
-        model (str): The name of the Hugging Face model.
-        revision (str, optional): The specific version
-        of the model to use. Defaults to 'main'.
+        model: The name of the Hugging Face model.
+        revision: The specific version of the model to use. 
+            Defaults to 'main'.
 
     Returns:
-        dict: A dictionary containing the pooling
-        type and whether normalization is used.
+        A dictionary containing the pooling type and whether 
+            normalization is used, or None if no pooling configuration is found.
     """
 
     modules_file_name = "modules.json"

vllm/transformers_utils/configs/jais.py

@@ -74,10 +74,10 @@ class JAISConfig(PretrainedConfig):
         use_cache (`bool`, *optional*, defaults to `True`):
             Whether or not the model should return the last key/values
             attentions (not used by all models).
-        scale_attn_by_inverse_layer_idx (`bool`, *optional*,
-            defaults to `False`):
-            Whether to additionally scale attention weights by
-            `1 / layer_idx + 1`.
+        scale_attn_by_inverse_layer_idx
+            (`bool`, *optional*, defaults to `False`): 
+            Whether to additionally scale attention weights 
+            by `1 / layer_idx + 1`.
         reorder_and_upcast_attn (`bool`, *optional*, defaults to `False`):
             Whether to scale keys (K) prior to computing attention
             (dot-product)

vllm/transformers_utils/configs/ultravox.py

@@ -37,10 +37,6 @@ class UltravoxConfig(transformers.PretrainedConfig):
             The initialization value for the layer normalization.
         projector_act (`str`, *optional*, defaults to `"swiglu"`):
             The activation function used by the multimodal projector.
-        text_model_lora_config (`LoraConfigSimplified`, *optional*):
-            The LoRA configuration for finetuning the text model.
-        audio_model_lora_config (`LoraConfigSimplified`, *optional*):
-            The LoRA configuration for finetuning the audio model.
         projector_ln_mid (`bool`, *optional*, defaults to `False`):
             Whether to apply layer normalization at the middle of the
             projector or at the end. Versions v0.4.1 and below

vllm/transformers_utils/processors/deepseek_vl2.py

@@ -25,6 +25,7 @@
 # CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 import math
+from typing import Any
 
 import torch
 import torchvision.transforms as T
@@ -178,17 +179,15 @@ class DeepseekVLV2Processor(ProcessorMixin):
         prompt: str,
         images: list[Image.Image],
         inference_mode: bool = True,
-        **kwargs,
+        **kwargs: Any,
     ):
         """
 
         Args:
             prompt (str): the formatted prompt;
-            conversations (list[dict]): conversations with a list of messages;
             images (list[ImageType]): the list of images;
             inference_mode (bool): if True, then remove the last eos token;
-            system_prompt (str): the system prompt;
-            **kwargs:
+            **kwargs: Additional keyword arguments.
 
         Returns:
             outputs (BaseProcessorOutput): the output of the processor,
@@ -259,7 +258,7 @@ class DeepseekVLV2Processor(ProcessorMixin):
         text: str,
         images: list[Image.Image],
         inference_mode: bool = True,
-        **kwargs,
+        **kwargs: Any,
     ):
         """
 

vllm/transformers_utils/runai_utils.py

@@ -33,7 +33,6 @@ def list_safetensors(path: str = "") -> list[str]:
 
     Args:
         path: The object storage path to list from.
-        allow_pattern: A list of patterns of which files to pull.
 
     Returns:
         list[str]: List of full object storage paths allowed by the pattern
@@ -54,8 +53,7 @@ class ObjectStorageModel:
         dir: The temporary created directory.
 
     Methods:
-        pull_files(): Pull model from object storage to the temporary
-        directory.
+        pull_files(): Pull model from object storage to the temporary directory.
     """
 
     def __init__(self) -> None:

vllm/transformers_utils/s3_utils.py

@@ -2,7 +2,7 @@
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import fnmatch
-from typing import Optional
+from typing import Any, Optional
 
 from vllm.utils import PlaceholderModule
 
@@ -26,7 +26,7 @@ def _filter_ignore(paths: list[str], patterns: list[str]) -> list[str]:
     ]
 
 
-def glob(s3=None,
+def glob(s3: Optional[Any] = None,
          path: str = "",
          allow_pattern: Optional[list[str]] = None) -> list[str]:
     """
@@ -51,7 +51,7 @@ def glob(s3=None,
 
 
 def list_files(
-        s3,
+        s3: Any,
         path: str,
         allow_pattern: Optional[list[str]] = None,
         ignore_pattern: Optional[list[str]] = None