Add vLLM provider

Signed-off-by: Yuan Tang <[email protected]>

.gitignore

@@ -13,3 +13,4 @@ xcuserdata/
 Package.resolved
 *.pte
 *.ipynb_checkpoints*
+.idea

llama_stack/distribution/templates/local-vllm-build.yaml

@@ -0,0 +1,10 @@
+name: local-vllm
+distribution_spec:
+  description: Use vLLM for running LLM inference
+  providers:
+    inference: remote::vllm
+    memory: meta-reference
+    safety: meta-reference
+    agents: meta-reference
+    telemetry: meta-reference
+image_type: conda

llama_stack/providers/adapters/inference/vllm/__init__.py

@@ -0,0 +1,24 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from .config import DatabricksImplConfig
+from .vllm import InferenceEndpointAdapter, VLLMAdapter
+
+
+async def get_adapter_impl(config: DatabricksImplConfig, _deps):
+    assert isinstance(config, DatabricksImplConfig), f"Unexpected config type: {type(config)}"
+
+    if config.url is not None:
+        impl = VLLMAdapter(config)
+    elif config.is_inference_endpoint():
+        impl = InferenceEndpointAdapter(config)
+    else:
+        raise ValueError(
+            "Invalid configuration. Specify either an URL or HF Inference Endpoint details (namespace and endpoint name)."
+        )
+
+    await impl.initialize()
+    return impl

llama_stack/providers/adapters/inference/vllm/config.py

@@ -0,0 +1,23 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from typing import Optional
+
+from llama_models.schema_utils import json_schema_type
+from pydantic import BaseModel, Field
+
+
+# TODO: Any other engine configs
+@json_schema_type
+class VLLMImplConfig(BaseModel):
+    url: Optional[str] = Field(
+        default=None,
+        description="The URL for the vLLM model serving endpoint",
+    )
+    api_token: Optional[str] = Field(
+        default=None,
+        description="The API token",
+    )

llama_stack/providers/adapters/inference/vllm/vllm.py

@@ -0,0 +1,262 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from typing import AsyncGenerator
+
+from llama_models.llama3.api.chat_format import ChatFormat
+
+from llama_models.llama3.api.datatypes import Message, StopReason
+from llama_models.llama3.api.tokenizer import Tokenizer
+from llama_models.sku_list import resolve_model
+
+from openai import OpenAI
+
+from llama_stack.apis.inference import *  # noqa: F403
+from llama_stack.providers.utils.inference.prepare_messages import prepare_messages
+
+from .config import VLLMImplConfig
+
+# TODO
+VLLM_SUPPORTED_MODELS = {}
+
+
+class VLLMInferenceAdapter(Inference):
+    def __init__(self, config: VLLMImplConfig) -> None:
+        self.config = config
+        tokenizer = Tokenizer.get_instance()
+        self.formatter = ChatFormat(tokenizer)
+
+    @property
+    def client(self) -> OpenAI:
+        return OpenAI(
+            api_key=self.config.api_token,
+            base_url=self.config.url
+        )
+
+    async def initialize(self) -> None:
+        return
+
+    async def shutdown(self) -> None:
+        pass
+
+    async def completion(self, request: CompletionRequest) -> AsyncGenerator:
+        raise NotImplementedError()
+
+    def _messages_to_vllm_messages(self, messages: list[Message]) -> list:
+        vllm_messages = []
+        for message in messages:
+            if message.role == "ipython":
+                role = "tool"
+            else:
+                role = message.role
+            vllm_messages.append({"role": role, "content": message.content})
+
+        return vllm_messages
+
+    def resolve_vllm_model(self, model_name: str) -> str:
+        model = resolve_model(model_name)
+        assert (
+            model is not None
+            and model.descriptor(shorten_default_variant=True)
+            in VLLM_SUPPORTED_MODELS
+        ), f"Unsupported model: {model_name}, use one of the supported models: {','.join(VLLM_SUPPORTED_MODELS.keys())}"
+
+        return VLLM_SUPPORTED_MODELS.get(
+            model.descriptor(shorten_default_variant=True)
+        )
+
+    def get_vllm_chat_options(self, request: ChatCompletionRequest) -> dict:
+        options = {}
+        # TODO
+        return options
+
+    async def chat_completion(
+        self,
+        model: str,
+        messages: List[Message],
+        sampling_params: Optional[SamplingParams] = SamplingParams(),
+        tools: Optional[List[ToolDefinition]] = None,
+        tool_choice: Optional[ToolChoice] = ToolChoice.auto,
+        tool_prompt_format: Optional[ToolPromptFormat] = ToolPromptFormat.json,
+        stream: Optional[bool] = False,
+        logprobs: Optional[LogProbConfig] = None,
+    ) -> AsyncGenerator:
+        # wrapper request to make it easier to pass around (internal only, not exposed to API)
+        request = ChatCompletionRequest(
+            model=model,
+            messages=messages,
+            sampling_params=sampling_params,
+            tools=tools or [],
+            tool_choice=tool_choice,
+            tool_prompt_format=tool_prompt_format,
+            stream=stream,
+            logprobs=logprobs,
+        )
+
+        # accumulate sampling params and other options to pass to vLLM
+        options = self.get_vllm_chat_options(request)
+        vllm_model = self.resolve_vllm_model(request.model)
+        messages = prepare_messages(request)
+        model_input = self.formatter.encode_dialog_prompt(messages)
+
+        input_tokens = len(model_input.tokens)
+        max_new_tokens = min(
+            request.sampling_params.max_tokens or (self.max_tokens - input_tokens),
+            self.max_tokens - input_tokens - 1,
+        )
+
+        print(f"Calculated max_new_tokens: {max_new_tokens}")
+
+        assert (
+            request.model == self.model_name
+        ), f"Model mismatch, expected {self.model_name}, got {request.model}"
+
+        if not request.stream:
+            r = self.client.chat.completions.create(
+                model=vllm_model,
+                messages=self._messages_to_vllm_messages(messages),
+                max_tokens=max_new_tokens,
+                stream=False,
+                **options,
+            )
+            stop_reason = None
+            if r.choices[0].finish_reason:
+                if (
+                    r.choices[0].finish_reason == "stop"
+                    or r.choices[0].finish_reason == "eos"
+                ):
+                    stop_reason = StopReason.end_of_turn
+                elif r.choices[0].finish_reason == "length":
+                    stop_reason = StopReason.out_of_tokens
+
+            completion_message = self.formatter.decode_assistant_message_from_content(
+                r.choices[0].message.content, stop_reason
+            )
+            yield ChatCompletionResponse(
+                completion_message=completion_message,
+                logprobs=None,
+            )
+        else:
+            yield ChatCompletionResponseStreamChunk(
+                event=ChatCompletionResponseEvent(
+                    event_type=ChatCompletionResponseEventType.start,
+                    delta="",
+                )
+            )
+
+            buffer = ""
+            ipython = False
+            stop_reason = None
+
+            for chunk in self.client.chat.completions.create(
+                model=vllm_model,
+                messages=self._messages_to_vllm_messages(messages),
+                max_tokens=max_new_tokens,
+                stream=True,
+                **options,
+            ):
+                if chunk.choices[0].finish_reason:
+                    if (
+                        stop_reason is None and chunk.choices[0].finish_reason == "stop"
+                    ) or (
+                        stop_reason is None and chunk.choices[0].finish_reason == "eos"
+                    ):
+                        stop_reason = StopReason.end_of_turn
+                    elif (
+                        stop_reason is None
+                        and chunk.choices[0].finish_reason == "length"
+                    ):
+                        stop_reason = StopReason.out_of_tokens
+                    break
+
+                text = chunk.choices[0].message.content
+                if text is None:
+                    continue
+
+                # check if it's a tool call ( aka starts with <|python_tag|> )
+                if not ipython and text.startswith("<|python_tag|>"):
+                    ipython = True
+                    yield ChatCompletionResponseStreamChunk(
+                        event=ChatCompletionResponseEvent(
+                            event_type=ChatCompletionResponseEventType.progress,
+                            delta=ToolCallDelta(
+                                content="",
+                                parse_status=ToolCallParseStatus.started,
+                            ),
+                        )
+                    )
+                    buffer += text
+                    continue
+
+                if ipython:
+                    if text == "<|eot_id|>":
+                        stop_reason = StopReason.end_of_turn
+                        text = ""
+                        continue
+                    elif text == "<|eom_id|>":
+                        stop_reason = StopReason.end_of_message
+                        text = ""
+                        continue
+
+                    buffer += text
+                    delta = ToolCallDelta(
+                        content=text,
+                        parse_status=ToolCallParseStatus.in_progress,
+                    )
+
+                    yield ChatCompletionResponseStreamChunk(
+                        event=ChatCompletionResponseEvent(
+                            event_type=ChatCompletionResponseEventType.progress,
+                            delta=delta,
+                            stop_reason=stop_reason,
+                        )
+                    )
+                else:
+                    buffer += text
+                    yield ChatCompletionResponseStreamChunk(
+                        event=ChatCompletionResponseEvent(
+                            event_type=ChatCompletionResponseEventType.progress,
+                            delta=text,
+                            stop_reason=stop_reason,
+                        )
+                    )
+
+            # parse tool calls and report errors
+            message = self.formatter.decode_assistant_message_from_content(
+                buffer, stop_reason
+            )
+            parsed_tool_calls = len(message.tool_calls) > 0
+            if ipython and not parsed_tool_calls:
+                yield ChatCompletionResponseStreamChunk(
+                    event=ChatCompletionResponseEvent(
+                        event_type=ChatCompletionResponseEventType.progress,
+                        delta=ToolCallDelta(
+                            content="",
+                            parse_status=ToolCallParseStatus.failure,
+                        ),
+                        stop_reason=stop_reason,
+                    )
+                )
+
+            for tool_call in message.tool_calls:
+                yield ChatCompletionResponseStreamChunk(
+                    event=ChatCompletionResponseEvent(
+                        event_type=ChatCompletionResponseEventType.progress,
+                        delta=ToolCallDelta(
+                            content=tool_call,
+                            parse_status=ToolCallParseStatus.success,
+                        ),
+                        stop_reason=stop_reason,
+                    )
+                )
+
+            yield ChatCompletionResponseStreamChunk(
+                event=ChatCompletionResponseEvent(
+                    event_type=ChatCompletionResponseEventType.complete,
+                    delta="",
+                    stop_reason=stop_reason,
+                )
+            )

llama_stack/providers/registry/inference.py

@@ -44,6 +44,14 @@ def available_providers() -> List[ProviderSpec]:
                 module="llama_stack.providers.adapters.inference.ollama",
             ),
         ),
+        remote_provider_spec(
+            api=Api.inference,
+            adapter=AdapterSpec(
+                adapter_type="vllm",
+                pip_packages=["openai"],
+                module="llama_stack.providers.adapters.inference.vllm",
+            ),
+        ),
         remote_provider_spec(
             api=Api.inference,
             adapter=AdapterSpec(