mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-06-27 18:50:41 +00:00
9436dd570d
# What does this PR do? We have support for embeddings in our Inference providers, but so far we haven't done the final step of actually registering the known embedding models and making sure they are extremely easy to use. This is one step towards that. ## Test Plan Run existing inference tests. ```bash $ cd llama_stack/providers/tests/inference $ pytest -s -v -k fireworks test_embeddings.py \ --inference-model nomic-ai/nomic-embed-text-v1.5 --env EMBEDDING_DIMENSION=784 $ pytest -s -v -k together test_embeddings.py \ --inference-model togethercomputer/m2-bert-80M-8k-retrieval --env EMBEDDING_DIMENSION=784 $ pytest -s -v -k ollama test_embeddings.py \ --inference-model all-minilm:latest --env EMBEDDING_DIMENSION=784 ``` The value of the EMBEDDING_DIMENSION isn't actually used in these tests, it is merely used by the test fixtures to check if the model is an LLM or Embedding.
103 lines
3 KiB
Python
103 lines
3 KiB
Python
# 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 llama_stack.apis.models.models import ModelType
|
|
from llama_stack.models.llama.datatypes import CoreModelId
|
|
from llama_stack.providers.utils.inference.model_registry import (
|
|
ProviderModelEntry,
|
|
build_hf_repo_model_entry,
|
|
build_model_entry,
|
|
)
|
|
|
|
model_entries = [
|
|
build_hf_repo_model_entry(
|
|
"llama3.1:8b-instruct-fp16",
|
|
CoreModelId.llama3_1_8b_instruct.value,
|
|
),
|
|
build_model_entry(
|
|
"llama3.1:8b",
|
|
CoreModelId.llama3_1_8b_instruct.value,
|
|
),
|
|
build_hf_repo_model_entry(
|
|
"llama3.1:70b-instruct-fp16",
|
|
CoreModelId.llama3_1_70b_instruct.value,
|
|
),
|
|
build_model_entry(
|
|
"llama3.1:70b",
|
|
CoreModelId.llama3_1_70b_instruct.value,
|
|
),
|
|
build_hf_repo_model_entry(
|
|
"llama3.1:405b-instruct-fp16",
|
|
CoreModelId.llama3_1_405b_instruct.value,
|
|
),
|
|
build_model_entry(
|
|
"llama3.1:405b",
|
|
CoreModelId.llama3_1_405b_instruct.value,
|
|
),
|
|
build_hf_repo_model_entry(
|
|
"llama3.2:1b-instruct-fp16",
|
|
CoreModelId.llama3_2_1b_instruct.value,
|
|
),
|
|
build_model_entry(
|
|
"llama3.2:1b",
|
|
CoreModelId.llama3_2_1b_instruct.value,
|
|
),
|
|
build_hf_repo_model_entry(
|
|
"llama3.2:3b-instruct-fp16",
|
|
CoreModelId.llama3_2_3b_instruct.value,
|
|
),
|
|
build_model_entry(
|
|
"llama3.2:3b",
|
|
CoreModelId.llama3_2_3b_instruct.value,
|
|
),
|
|
build_hf_repo_model_entry(
|
|
"llama3.2-vision:11b-instruct-fp16",
|
|
CoreModelId.llama3_2_11b_vision_instruct.value,
|
|
),
|
|
build_model_entry(
|
|
"llama3.2-vision:latest",
|
|
CoreModelId.llama3_2_11b_vision_instruct.value,
|
|
),
|
|
build_hf_repo_model_entry(
|
|
"llama3.2-vision:90b-instruct-fp16",
|
|
CoreModelId.llama3_2_90b_vision_instruct.value,
|
|
),
|
|
build_model_entry(
|
|
"llama3.2-vision:90b",
|
|
CoreModelId.llama3_2_90b_vision_instruct.value,
|
|
),
|
|
build_hf_repo_model_entry(
|
|
"llama3.3:70b",
|
|
CoreModelId.llama3_3_70b_instruct.value,
|
|
),
|
|
# The Llama Guard models don't have their full fp16 versions
|
|
# so we are going to alias their default version to the canonical SKU
|
|
build_hf_repo_model_entry(
|
|
"llama-guard3:8b",
|
|
CoreModelId.llama_guard_3_8b.value,
|
|
),
|
|
build_hf_repo_model_entry(
|
|
"llama-guard3:1b",
|
|
CoreModelId.llama_guard_3_1b.value,
|
|
),
|
|
ProviderModelEntry(
|
|
provider_model_id="all-minilm:latest",
|
|
aliases=["all-minilm"],
|
|
model_type=ModelType.embedding,
|
|
metadata={
|
|
"embedding_dimensions": 384,
|
|
"context_length": 512,
|
|
},
|
|
),
|
|
ProviderModelEntry(
|
|
provider_model_id="nomic-embed-text",
|
|
model_type=ModelType.embedding,
|
|
metadata={
|
|
"embedding_dimensions": 768,
|
|
"context_length": 8192,
|
|
},
|
|
),
|
|
]
|