phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-08-01 16:24:44 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

support adding alias for models without hf repo/sku entry

Browse source
This commit is contained in:
Dinesh Yeduguru 2024-11-18 23:31:04 -08:00
parent fcc2132e6f
commit 8bd0a33206
2 changed files with 18 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

15
llama_stack/providers/remote/inference/ollama/ollama.py
View file

@ -16,6 +16,7 @@ from ollama import AsyncClient
from llama_stack.providers.utils.inference.model_registry import (
build_model_alias,
build_model_alias_with_just_llama_model,
ModelRegistryHelper,
)
@ -44,7 +45,7 @@ model_aliases = [
"llama3.1:8b-instruct-fp16",
CoreModelId.llama3_1_8b_instruct.value,
),
build_model_alias(
build_model_alias_with_just_llama_model(
"llama3.1:8b",
CoreModelId.llama3_1_8b_instruct.value,
),
@ -52,7 +53,7 @@ model_aliases = [
"llama3.1:70b-instruct-fp16",
CoreModelId.llama3_1_70b_instruct.value,
),
build_model_alias(
build_model_alias_with_just_llama_model(
"llama3.1:70b",
CoreModelId.llama3_1_70b_instruct.value,
),
@ -64,19 +65,19 @@ model_aliases = [
"llama3.2:3b-instruct-fp16",
CoreModelId.llama3_2_3b_instruct.value,
),
build_model_alias(
build_model_alias_with_just_llama_model(
"llama3.2:1b",
CoreModelId.llama3_2_1b_instruct.value,
),
build_model_alias(
build_model_alias_with_just_llama_model(
"llama3.2:3b",
CoreModelId.llama3_2_3b_instruct.value,
),
build_model_alias(
build_model_alias_with_just_llama_model(
"llama-guard3:8b",
CoreModelId.llama_guard_3_8b.value,
),
build_model_alias(
build_model_alias_with_just_llama_model(
"llama-guard3:1b",
CoreModelId.llama_guard_3_1b.value,
),
@ -84,7 +85,7 @@ model_aliases = [
"x/llama3.2-vision:11b-instruct-fp16",
CoreModelId.llama3_2_11b_vision_instruct.value,
),
build_model_alias(
build_model_alias_with_just_llama_model(
"llama3.2-vision",
CoreModelId.llama3_2_11b_vision_instruct.value,
),

10
llama_stack/providers/utils/inference/model_registry.py
View file

@ -36,6 +36,16 @@ def build_model_alias(provider_model_id: str, model_descriptor: str) -> ModelAli
)
def build_model_alias_with_just_llama_model(
provider_model_id: str, model_descriptor: str
) -> ModelAlias:
return ModelAlias(
provider_model_id=provider_model_id,
aliases=[],
llama_model=model_descriptor,
)
class ModelRegistryHelper(ModelsProtocolPrivate):
def __init__(self, model_aliases: List[ModelAlias]):
self.alias_to_provider_id_map = {}