forked from phoenix-oss/llama-stack-mirror
Remove "routing_table" and "routing_key" concepts for the user (#201)
This PR makes several core changes to the developer experience surrounding Llama Stack. Background: PR #92 introduced the notion of "routing" to the Llama Stack. It introduces three object types: (1) models, (2) shields and (3) memory banks. Each of these objects can be associated with a distinct provider. So you can get model A to be inferenced locally while model B, C can be inference remotely (e.g.) However, this had a few drawbacks: you could not address the provider instances -- i.e., if you configured "meta-reference" with a given model, you could not assign an identifier to this instance which you could re-use later. the above meant that you could not register a "routing_key" (e.g. model) dynamically and say "please use this existing provider I have already configured" for a new model. the terms "routing_table" and "routing_key" were exposed directly to the user. in my view, this is way too much overhead for a new user (which almost everyone is.) people come to the stack wanting to do ML and encounter a completely unexpected term. What this PR does: This PR structures the run config with only a single prominent key: - providers Providers are instances of configured provider types. Here's an example which shows two instances of the remote::tgi provider which are serving two different models. providers: inference: - provider_id: foo provider_type: remote::tgi config: { ... } - provider_id: bar provider_type: remote::tgi config: { ... } Secondly, the PR adds dynamic registration of { models | shields | memory_banks } to the API surface. The distribution still acts like a "routing table" (as previously) except that it asks the backing providers for a listing of these objects. For example it asks a TGI or Ollama inference adapter what models it is serving. Only the models that are being actually served can be requested by the user for inference. Otherwise, the Stack server will throw an error. When dynamically registering these objects, you can use the provider IDs shown above. Info about providers can be obtained using the Api.inspect set of endpoints (/providers, /routes, etc.) The above examples shows the correspondence between inference providers and models registry items. Things work similarly for the safety <=> shields and memory <=> memory_banks pairs. Registry: This PR also makes it so that Providers need to implement additional methods for registering and listing objects. For example, each Inference provider is now expected to implement the ModelsProtocolPrivate protocol (naming is not great!) which consists of two methods register_model list_models The goal is to inform the provider that a certain model needs to be supported so the provider can make any relevant backend changes if needed (or throw an error if the model cannot be supported.) There are many other cleanups included some of which are detailed in a follow-up comment.
This commit is contained in:
Ashwin Bharambe
GitHub
parent
8c3010553f
commit
6bb57e72a7
93 changed files with 4697 additions and 4457 deletions
|
|
@ -8,17 +8,22 @@ 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.datatypes import Message
|
||||
from llama_models.llama3.api.tokenizer import Tokenizer
|
||||
|
||||
from together import Together
|
||||
|
||||
from llama_stack.apis.inference import * # noqa: F403
|
||||
from llama_stack.distribution.request_headers import NeedsRequestProviderData
|
||||
from llama_stack.providers.utils.inference.augment_messages import (
|
||||
augment_messages_for_tools,
|
||||
from llama_stack.providers.utils.inference.model_registry import ModelRegistryHelper
|
||||
from llama_stack.providers.utils.inference.openai_compat import (
|
||||
get_sampling_options,
|
||||
process_chat_completion_response,
|
||||
process_chat_completion_stream_response,
|
||||
)
|
||||
from llama_stack.providers.utils.inference.prompt_adapter import (
|
||||
chat_completion_request_to_prompt,
|
||||
)
|
||||
from llama_stack.providers.utils.inference.routable import RoutableProviderForModels
|
||||
|
||||
from .config import TogetherImplConfig
|
||||
|
||||
|
|
@ -34,19 +39,14 @@ TOGETHER_SUPPORTED_MODELS = {
|
|||
|
||||
|
||||
class TogetherInferenceAdapter(
|
||||
Inference, NeedsRequestProviderData, RoutableProviderForModels
|
||||
ModelRegistryHelper, Inference, NeedsRequestProviderData
|
||||
):
|
||||
def __init__(self, config: TogetherImplConfig) -> None:
|
||||
RoutableProviderForModels.__init__(
|
||||
ModelRegistryHelper.__init__(
|
||||
self, stack_to_provider_models_map=TOGETHER_SUPPORTED_MODELS
|
||||
)
|
||||
self.config = config
|
||||
tokenizer = Tokenizer.get_instance()
|
||||
self.formatter = ChatFormat(tokenizer)
|
||||
|
||||
@property
|
||||
def client(self) -> Together:
|
||||
return Together(api_key=self.config.api_key)
|
||||
self.formatter = ChatFormat(Tokenizer.get_instance())
|
||||
|
||||
async def initialize(self) -> None:
|
||||
return
|
||||
|
|
@ -64,27 +64,7 @@ class TogetherInferenceAdapter(
|
|||
) -> AsyncGenerator:
|
||||
raise NotImplementedError()
|
||||
|
||||
def _messages_to_together_messages(self, messages: list[Message]) -> list:
|
||||
together_messages = []
|
||||
for message in messages:
|
||||
if message.role == "ipython":
|
||||
role = "tool"
|
||||
else:
|
||||
role = message.role
|
||||
together_messages.append({"role": role, "content": message.content})
|
||||
|
||||
return together_messages
|
||||
|
||||
def get_together_chat_options(self, request: ChatCompletionRequest) -> dict:
|
||||
options = {}
|
||||
if request.sampling_params is not None:
|
||||
for attr in {"temperature", "top_p", "top_k", "max_tokens"}:
|
||||
if getattr(request.sampling_params, attr):
|
||||
options[attr] = getattr(request.sampling_params, attr)
|
||||
|
||||
return options
|
||||
|
||||
async def chat_completion(
|
||||
def chat_completion(
|
||||
self,
|
||||
model: str,
|
||||
messages: List[Message],
|
||||
|
|
@ -95,7 +75,6 @@ class TogetherInferenceAdapter(
|
|||
stream: Optional[bool] = False,
|
||||
logprobs: Optional[LogProbConfig] = None,
|
||||
) -> AsyncGenerator:
|
||||
|
||||
together_api_key = None
|
||||
if self.config.api_key is not None:
|
||||
together_api_key = self.config.api_key
|
||||
|
|
@ -108,7 +87,6 @@ class TogetherInferenceAdapter(
|
|||
together_api_key = provider_data.together_api_key
|
||||
|
||||
client = Together(api_key=together_api_key)
|
||||
# wrapper request to make it easier to pass around (internal only, not exposed to API)
|
||||
request = ChatCompletionRequest(
|
||||
model=model,
|
||||
messages=messages,
|
||||
|
|
@ -120,146 +98,46 @@ class TogetherInferenceAdapter(
|
|||
logprobs=logprobs,
|
||||
)
|
||||
|
||||
# accumulate sampling params and other options to pass to together
|
||||
options = self.get_together_chat_options(request)
|
||||
together_model = self.map_to_provider_model(request.model)
|
||||
messages = augment_messages_for_tools(request)
|
||||
|
||||
if not request.stream:
|
||||
# TODO: might need to add back an async here
|
||||
r = client.chat.completions.create(
|
||||
model=together_model,
|
||||
messages=self._messages_to_together_messages(messages),
|
||||
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,
|
||||
)
|
||||
if stream:
|
||||
return self._stream_chat_completion(request, client)
|
||||
else:
|
||||
yield ChatCompletionResponseStreamChunk(
|
||||
event=ChatCompletionResponseEvent(
|
||||
event_type=ChatCompletionResponseEventType.start,
|
||||
delta="",
|
||||
)
|
||||
)
|
||||
return self._nonstream_chat_completion(request, client)
|
||||
|
||||
buffer = ""
|
||||
ipython = False
|
||||
stop_reason = None
|
||||
async def _nonstream_chat_completion(
|
||||
self, request: ChatCompletionRequest, client: Together
|
||||
) -> ChatCompletionResponse:
|
||||
params = self._get_params(request)
|
||||
r = client.completions.create(**params)
|
||||
return process_chat_completion_response(request, r, self.formatter)
|
||||
|
||||
for chunk in client.chat.completions.create(
|
||||
model=together_model,
|
||||
messages=self._messages_to_together_messages(messages),
|
||||
stream=True,
|
||||
**options,
|
||||
):
|
||||
if finish_reason := chunk.choices[0].finish_reason:
|
||||
if stop_reason is None and finish_reason in ["stop", "eos"]:
|
||||
stop_reason = StopReason.end_of_turn
|
||||
elif stop_reason is None and finish_reason == "length":
|
||||
stop_reason = StopReason.out_of_tokens
|
||||
break
|
||||
async def _stream_chat_completion(
|
||||
self, request: ChatCompletionRequest, client: Together
|
||||
) -> AsyncGenerator:
|
||||
params = self._get_params(request)
|
||||
|
||||
text = chunk.choices[0].delta.content
|
||||
if text is None:
|
||||
continue
|
||||
# if we shift to TogetherAsyncClient, we won't need this wrapper
|
||||
async def _to_async_generator():
|
||||
s = client.completions.create(**params)
|
||||
for chunk in s:
|
||||
yield chunk
|
||||
|
||||
# check if its 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
|
||||
stream = _to_async_generator()
|
||||
async for chunk in process_chat_completion_stream_response(
|
||||
request, stream, self.formatter
|
||||
):
|
||||
yield chunk
|
||||
|
||||
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
|
||||
def _get_params(self, request: ChatCompletionRequest) -> dict:
|
||||
return {
|
||||
"model": self.map_to_provider_model(request.model),
|
||||
"prompt": chat_completion_request_to_prompt(request, self.formatter),
|
||||
"stream": request.stream,
|
||||
**get_sampling_options(request),
|
||||
}
|
||||
|
||||
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,
|
||||
)
|
||||
)
|
||||
async def embeddings(
|
||||
self,
|
||||
model: str,
|
||||
contents: List[InterleavedTextMedia],
|
||||
) -> EmbeddingsResponse:
|
||||
raise NotImplementedError()
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue