phoenix-oss/llama-stack
1
0
Fork 0
forked from phoenix-oss/llama-stack-mirror
Code Pull requests Releases Packages 2 Activity Actions

chore: remove llama_models.llama3.api imports from providers (#1107)

Browse source 
There should be a choke-point for llama3.api imports -- this is the
prompt adapter. Creating a ChatFormat() object on demand is inexpensive.
The underlying Tokenizer is a singleton anyway.
This commit is contained in:
Ashwin Bharambe 2025-02-19 19:01:29 -08:00 • committed by GitHub
parent e9b8259cf9
commit cdcbeb005b
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
13 changed files with 77 additions and 113 deletions
Download patch file Download diff file Expand all files Collapse all files

9
llama_stack/providers/remote/inference/bedrock/bedrock.py
View file

@ -8,8 +8,6 @@ import json
from typing import AsyncGenerator, AsyncIterator, Dict, List, Optional, Union
from botocore.client import BaseClient
from llama_models.llama3.api.chat_format import ChatFormat
from llama_models.llama3.api.tokenizer import Tokenizer
from llama_stack.apis.common.content_types import InterleavedContent
from llama_stack.apis.inference import (
@ -54,7 +52,6 @@ class BedrockInferenceAdapter(ModelRegistryHelper, Inference):
self._config = config
self._client = create_bedrock_client(config)
self.formatter = ChatFormat(Tokenizer.get_instance())
@property
def client(self) -> BaseClient:
@ -119,7 +116,7 @@ class BedrockInferenceAdapter(ModelRegistryHelper, Inference):
)
response = OpenAICompatCompletionResponse(choices=[choice])
return process_chat_completion_response(response, self.formatter, request)
return process_chat_completion_response(response, request)
async def _stream_chat_completion(self, request: ChatCompletionRequest) -> AsyncGenerator:
params = await self._get_params_for_chat_completion(request)
@ -137,7 +134,7 @@ class BedrockInferenceAdapter(ModelRegistryHelper, Inference):
yield OpenAICompatCompletionResponse(choices=[choice])
stream = _generate_and_convert_to_openai_compat()
async for chunk in process_chat_completion_stream_response(stream, self.formatter, request):
async for chunk in process_chat_completion_stream_response(stream, request):
yield chunk
async def _get_params_for_chat_completion(self, request: ChatCompletionRequest) -> Dict:
@ -151,7 +148,7 @@ class BedrockInferenceAdapter(ModelRegistryHelper, Inference):
if sampling_params.repetition_penalty > 0:
options["repetition_penalty"] = sampling_params.repetition_penalty
prompt = await chat_completion_request_to_prompt(request, self.get_llama_model(request.model), self.formatter)
prompt = await chat_completion_request_to_prompt(request, self.get_llama_model(request.model))
return {
"modelId": bedrock_model,
"body": json.dumps(