[VLM] Support caching in merged multi-modal processor (#11396)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

docs/source/conf.py

@@ -191,6 +191,7 @@ def linkcode_resolve(domain, info):
 
 # Mock out external dependencies here, otherwise the autodoc pages may be blank.
 autodoc_mock_imports = [
+    "blake3",
     "compressed_tensors",
     "cpuinfo",
     "cv2",
@@ -207,7 +208,7 @@ autodoc_mock_imports = [
     "tensorizer",
     "pynvml",
     "outlines",
-    "xgrammar,"
+    "xgrammar",
     "librosa",
     "soundfile",
     "gguf",

docs/source/design/multimodal/multimodal_index.md

@@ -45,31 +45,23 @@ adding_multimodal_plugin
 ### Base Classes
 
 ```{eval-rst}
-.. autodata:: vllm.multimodal.NestedTensors
-```
-
-```{eval-rst}
-.. autodata:: vllm.multimodal.BatchedTensorInputs
-```
-
-```{eval-rst}
-.. autoclass:: vllm.multimodal.MultiModalDataBuiltins
+.. automodule:: vllm.multimodal.base
     :members:
     :show-inheritance:
 ```
 
-```{eval-rst}
-.. autodata:: vllm.multimodal.MultiModalDataDict
-```
+### Input Classes
 
 ```{eval-rst}
-.. autoclass:: vllm.multimodal.MultiModalKwargs
+.. automodule:: vllm.multimodal.inputs
     :members:
     :show-inheritance:
 ```
 
+### Audio Classes
+
 ```{eval-rst}
-.. autoclass:: vllm.multimodal.MultiModalPlugin
+.. automodule:: vllm.multimodal.audio
     :members:
     :show-inheritance:
 ```
@@ -81,3 +73,11 @@ adding_multimodal_plugin
     :members:
     :show-inheritance:
 ```
+
+### Video Classes
+
+```{eval-rst}
+.. automodule:: vllm.multimodal.video
+    :members:
+    :show-inheritance:
+```

docs/source/models/supported_models.md

@@ -755,8 +755,7 @@ vLLM currently only supports adding LoRA to the language backbone of multimodal
 ```
 
 ```{note}
-To use {code}`TIGER-Lab/Mantis-8B-siglip-llama3`, you have to install their GitHub repo ({code}`pip install git+https://github.com/TIGER-AI-Lab/Mantis.git`)
-and pass {code}`--hf_overrides '{"architectures": ["MantisForConditionalGeneration"]}'` when running vLLM.
+To use {code}`TIGER-Lab/Mantis-8B-siglip-llama3`, you have pass {code}`--hf_overrides '{"architectures": ["MantisForConditionalGeneration"]}'` when running vLLM.
 ```
 
 ```{note}