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

11
llama_stack/providers/remote/inference/sambanova/sambanova.py
View file

@ -7,8 +7,6 @@
import json
from typing import AsyncGenerator
from llama_models.llama3.api.chat_format import ChatFormat
from llama_models.llama3.api.tokenizer import Tokenizer
from openai import OpenAI
from llama_stack.apis.common.content_types import (
@ -38,13 +36,8 @@ from .models import MODEL_ALIASES
class SambaNovaInferenceAdapter(ModelRegistryHelper, Inference):
def __init__(self, config: SambaNovaImplConfig) -> None:
ModelRegistryHelper.__init__(
self,
model_aliases=MODEL_ALIASES,
)
ModelRegistryHelper.__init__(self, model_aliases=MODEL_ALIASES)
self.config = config
self.formatter = ChatFormat(Tokenizer.get_instance())
async def initialize(self) -> None:
return
@ -120,7 +113,7 @@ class SambaNovaInferenceAdapter(ModelRegistryHelper, Inference):
yield chunk
stream = _to_async_generator()
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 embeddings(