[Frontend] Chat-based Embeddings API (vllm-project#9759)

docs/requirements-docs.txt

@@ -13,5 +13,7 @@ torch
 py-cpuinfo
 transformers
 mistral_common >= 1.3.4
+aiohttp
+starlette
 openai # Required by docs/source/serving/openai_compatible_server.md's vllm.entrypoints.openai.cli_args
 partial-json-parser # Required by docs/source/serving/openai_compatible_server.md's vllm.entrypoints.openai.cli_args

docs/source/conf.py

@@ -96,7 +96,6 @@ def setup(app):
 
 # Mock out external dependencies here, otherwise the autodoc pages may be blank.
 autodoc_mock_imports = [
-    "aiohttp",
     "compressed_tensors",
     "cpuinfo",
     "cv2",
@@ -143,6 +142,7 @@ def add_line(self, line: str, source: str, *lineno: int) -> None:
     "python": ("https://docs.python.org/3", None),
     "typing_extensions":
     ("https://typing-extensions.readthedocs.io/en/latest", None),
+    "aiohttp": ("https://docs.aiohttp.org/en/stable", None),
     "pillow": ("https://pillow.readthedocs.io/en/stable", None),
     "numpy": ("https://numpy.org/doc/stable", None),
     "torch": ("https://pytorch.org/docs/stable", None),

docs/source/dev/pooling_params.rst

@@ -0,0 +1,5 @@
+Pooling Parameters
+==================
+
+.. autoclass:: vllm.PoolingParams
+    :members:

docs/source/getting_started/quickstart.rst

@@ -138,10 +138,10 @@ Since this server is compatible with OpenAI API, you can use it as a drop-in rep
 
 A more detailed client example can be found `here <https://github.com/vllm-project/vllm/blob/main/examples/openai_completion_client.py>`__.
 
-OpenAI Chat API with vLLM
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+OpenAI Chat Completions API with vLLM
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-vLLM is designed to also support the OpenAI Chat API. The chat interface is a more dynamic, interactive way to communicate with the model, allowing back-and-forth exchanges that can be stored in the chat history. This is useful for tasks that require context or more detailed explanations.
+vLLM is designed to also support the OpenAI Chat Completions API. The chat interface is a more dynamic, interactive way to communicate with the model, allowing back-and-forth exchanges that can be stored in the chat history. This is useful for tasks that require context or more detailed explanations.
 
 You can use the `create chat completion <https://platform.openai.com/docs/api-reference/chat/completions/create>`_ endpoint to interact with the model:
 
@@ -157,7 +157,7 @@ You can use the `create chat completion <https://platform.openai.com/docs/api-re
     $         ]
     $     }'
 
-Alternatively, you can use the `openai` python package:
+Alternatively, you can use the ``openai`` python package:
 
 .. code-block:: python
 

docs/source/index.rst

@@ -134,6 +134,7 @@ Documentation
    :caption: Developer Documentation
 
    dev/sampling_params
+   dev/pooling_params
    dev/offline_inference/offline_index
    dev/engine/engine_index
    dev/kernel/paged_attention

docs/source/models/vlm.rst

@@ -185,7 +185,7 @@ Below is an example on how to launch the same ``microsoft/Phi-3.5-vision-instruc
       --trust-remote-code --max-model-len 4096 --limit-mm-per-prompt image=2
 
 .. important::
-    Since OpenAI Vision API is based on `Chat Completions <https://platform.openai.com/docs/api-reference/chat>`_ API,
+    Since OpenAI Vision API is based on `Chat Completions API <https://platform.openai.com/docs/api-reference/chat>`_,
     a chat template is **required** to launch the API server.
 
     Although Phi-3.5-Vision comes with a chat template, for other models you may have to provide one if the model's tokenizer does not come with it.
@@ -243,6 +243,10 @@ To consume the server, you can use the OpenAI client like in the example below:
 
 A full code example can be found in `examples/openai_api_client_for_multimodal.py <https://github.com/vllm-project/vllm/blob/main/examples/openai_api_client_for_multimodal.py>`_.
 
+.. tip::
+    There is no need to place image placeholders in the text content of the API request - they are already represented by the image content.
+    In fact, you can place image placeholders in the middle of the text by interleaving text and image content.
+
 .. note::
 
     By default, the timeout for fetching images through http url is ``5`` seconds. You can override this by setting the environment variable:
@@ -251,5 +255,49 @@ A full code example can be found in `examples/openai_api_client_for_multimodal.p
 
         $ export VLLM_IMAGE_FETCH_TIMEOUT=<timeout>
 
-.. note::
-    There is no need to format the prompt in the API request since it will be handled by the server.
+Chat Embeddings API
+^^^^^^^^^^^^^^^^^^^
+
+vLLM's Chat Embeddings API is a superset of OpenAI's `Embeddings API <https://platform.openai.com/docs/api-reference/embeddings>`_,
+where a list of ``messages`` can be passed instead of batched ``inputs``. This enables multi-modal inputs to be passed to embedding models.
+
+.. tip::
+    The schema of ``messages`` is exactly the same as in Chat Completions API.
+
+In this example, we will serve the ``TIGER-Lab/VLM2Vec-Full`` model.
+
+.. code-block:: bash
+
+    vllm serve TIGER-Lab/VLM2Vec-Full --task embedding \
+      --trust-remote-code --max-model-len 4096
+
+.. important::
+
+    Since VLM2Vec has the same model architecture as Phi-3.5-Vision, we have to explicitly pass ``--task embedding``
+    to run this model in embedding mode instead of text generation mode.
+
+Since this schema is not defined by OpenAI client, we post a request to the server using the lower-level ``requests`` library:
+
+.. code-block:: python
+
+    import requests
+
+    image_url = "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
+
+    response = requests.post(
+        "http://localhost:8000/v1/embeddings",
+        json={
+            "model": "TIGER-Lab/VLM2Vec-Full",
+            "messages": [{
+                "role": "user",
+                "content": [
+                    {"type": "image_url", "image_url": {"url": image_url}},
+                    {"type": "text", "text": "Represent the given image."},
+                ],
+            }],
+            "encoding_format": "float",
+        },
+    )
+    response.raise_for_status()
+    response_json = response.json()
+    print("Embedding output:", response_json["data"][0]["embedding"])

docs/source/serving/openai_compatible_server.md

@@ -26,13 +26,26 @@ print(completion.choices[0].message)
 ```
 
 ## API Reference
-Please see the [OpenAI API Reference](https://platform.openai.com/docs/api-reference) for more information on the API. We support all parameters except:
-- Chat: `tools`, and `tool_choice`.
-- Completions: `suffix`.
 
-vLLM also provides experimental support for OpenAI Vision API compatible inference. See more details in [Using VLMs](../models/vlm.rst).
+We currently support the following OpenAI APIs:
+
+- [Completions API](https://platform.openai.com/docs/api-reference/completions)
+  - *Note: `suffix` parameter is not supported.*
+- [Chat Completions API](https://platform.openai.com/docs/api-reference/chat)
+  - [Vision](https://platform.openai.com/docs/guides/vision)-related parameters are supported; see [Using VLMs](../models/vlm.rst).
+    - *Note: `image_url.detail` parameter is not supported.*
+  - We also support `audio_url` content type for audio files.
+    - Refer to [vllm.entrypoints.chat_utils](https://github.com/vllm-project/vllm/tree/main/vllm/entrypoints/chat_utils.py) for the exact schema.
+    - *TODO: Support `input_audio` content type as defined [here](https://github.com/openai/openai-python/blob/v1.52.2/src/openai/types/chat/chat_completion_content_part_input_audio_param.py).*
+  - *Note: `parallel_tool_calls` and `user` parameters are ignored.*
+- [Embeddings API](https://platform.openai.com/docs/api-reference/embeddings)
+  - Instead of `inputs`, you can pass in a list of `messages` (same schema as Chat Completions API),
+    which will be treated as a single prompt to the model according to its chat template.
+    - This enables multi-modal inputs to be passed to embedding models, see [Using VLMs](../models/vlm.rst).
+  - *Note: You should run `vllm serve` with `--task embedding` to ensure that the model is being run in embedding mode.*
 
 ## Extra Parameters
+
 vLLM supports a set of parameters that are not part of the OpenAI API.
 In order to use them, you can pass them as extra parameters in the OpenAI client.
 Or directly merge them into the JSON payload if you are using HTTP call directly.
@@ -49,7 +62,26 @@ completion = client.chat.completions.create(
 )
 ```
 
-### Extra Parameters for Chat API
+### Extra Parameters for Completions API
+
+The following [sampling parameters (click through to see documentation)](../dev/sampling_params.rst) are supported.
+
+```{literalinclude} ../../../vllm/entrypoints/openai/protocol.py
+:language: python
+:start-after: begin-completion-sampling-params
+:end-before: end-completion-sampling-params
+```
+
+The following extra parameters are supported:
+
+```{literalinclude} ../../../vllm/entrypoints/openai/protocol.py
+:language: python
+:start-after: begin-completion-extra-params
+:end-before: end-completion-extra-params
+```
+
+### Extra Parameters for Chat Completions API
+
 The following [sampling parameters (click through to see documentation)](../dev/sampling_params.rst) are supported.
 
 ```{literalinclude} ../../../vllm/entrypoints/openai/protocol.py
@@ -66,21 +98,22 @@ The following extra parameters are supported:
 :end-before: end-chat-completion-extra-params
 ```
 
-### Extra Parameters for Completions API
-The following [sampling parameters (click through to see documentation)](../dev/sampling_params.rst) are supported.
+### Extra Parameters for Embeddings API
+
+The following [pooling parameters (click through to see documentation)](../dev/pooling_params.rst) are supported.
 
 ```{literalinclude} ../../../vllm/entrypoints/openai/protocol.py
 :language: python
-:start-after: begin-completion-sampling-params
-:end-before: end-completion-sampling-params
+:start-after: begin-embedding-pooling-params
+:end-before: end-embedding-pooling-params
 ```
 
 The following extra parameters are supported:
 
 ```{literalinclude} ../../../vllm/entrypoints/openai/protocol.py
 :language: python
-:start-after: begin-completion-extra-params
-:end-before: end-completion-extra-params
+:start-after: begin-embedding-extra-params
+:end-before: end-embedding-extra-params
 ```
 
 ## Chat Template

tests/entrypoints/openai/test_basic.py

@@ -1,7 +1,6 @@
 from http import HTTPStatus
 from typing import List
 
-import openai
 import pytest
 import pytest_asyncio
 import requests
@@ -83,10 +82,8 @@ async def client(server):
     indirect=True,
 )
 @pytest.mark.asyncio
-async def test_show_version(client: openai.AsyncOpenAI):
-    base_url = str(client.base_url)[:-3].strip("/")
-
-    response = requests.get(base_url + "/version")
+async def test_show_version(server: RemoteOpenAIServer):
+    response = requests.get(server.url_for("version"))
     response.raise_for_status()
 
     assert response.json() == {"version": VLLM_VERSION}
@@ -102,9 +99,7 @@ async def test_show_version(client: openai.AsyncOpenAI):
     indirect=True,
 )
 @pytest.mark.asyncio
-async def test_check_health(client: openai.AsyncOpenAI):
-    base_url = str(client.base_url)[:-3].strip("/")
-
-    response = requests.get(base_url + "/health")
+async def test_check_health(server: RemoteOpenAIServer):
+    response = requests.get(server.url_for("health"))
 
     assert response.status_code == HTTPStatus.OK