mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-06-27 18:50:41 +00:00
63e6acd0c3
# What does this PR do?
This PR introduces more non-llama model support to llama stack.
Providers introduced: openai, anthropic and gemini. All of these
providers use essentially the same piece of code -- the implementation
works via the `litellm` library.
We will expose only specific models for providers we enable making sure
they all work well and pass tests. This setup (instead of automatically
enabling _all_ providers and models allowed by LiteLLM) ensures we can
also perform any needed prompt tuning on a per-model basis as needed
(just like we do it for llama models.)
## Test Plan
```bash
#!/bin/bash
args=("$@")
for model in openai/gpt-4o anthropic/claude-3-5-sonnet-latest gemini/gemini-1.5-flash; do
LLAMA_STACK_CONFIG=dev pytest -s -v tests/client-sdk/inference/test_text_inference.py \
--embedding-model=all-MiniLM-L6-v2 \
--vision-inference-model="" \
--inference-model=$model "${args[@]}"
done
```
258 lines
9.9 KiB
Python
258 lines
9.9 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 typing import List
|
|
|
|
from llama_stack.providers.datatypes import (
|
|
AdapterSpec,
|
|
Api,
|
|
InlineProviderSpec,
|
|
ProviderSpec,
|
|
remote_provider_spec,
|
|
)
|
|
|
|
META_REFERENCE_DEPS = [
|
|
"accelerate",
|
|
"blobfile",
|
|
"fairscale",
|
|
"torch",
|
|
"torchvision",
|
|
"transformers",
|
|
"zmq",
|
|
"lm-format-enforcer",
|
|
"sentence-transformers",
|
|
]
|
|
|
|
|
|
def available_providers() -> List[ProviderSpec]:
|
|
return [
|
|
InlineProviderSpec(
|
|
api=Api.inference,
|
|
provider_type="inline::meta-reference",
|
|
pip_packages=META_REFERENCE_DEPS,
|
|
module="llama_stack.providers.inline.inference.meta_reference",
|
|
config_class="llama_stack.providers.inline.inference.meta_reference.MetaReferenceInferenceConfig",
|
|
),
|
|
InlineProviderSpec(
|
|
api=Api.inference,
|
|
provider_type="inline::meta-reference-quantized",
|
|
pip_packages=(
|
|
META_REFERENCE_DEPS
|
|
+ [
|
|
"fbgemm-gpu",
|
|
"torchao==0.5.0",
|
|
]
|
|
),
|
|
module="llama_stack.providers.inline.inference.meta_reference",
|
|
config_class="llama_stack.providers.inline.inference.meta_reference.MetaReferenceQuantizedInferenceConfig",
|
|
),
|
|
InlineProviderSpec(
|
|
api=Api.inference,
|
|
provider_type="inline::vllm",
|
|
pip_packages=[
|
|
"vllm",
|
|
],
|
|
module="llama_stack.providers.inline.inference.vllm",
|
|
config_class="llama_stack.providers.inline.inference.vllm.VLLMConfig",
|
|
),
|
|
InlineProviderSpec(
|
|
api=Api.inference,
|
|
provider_type="inline::sentence-transformers",
|
|
pip_packages=[
|
|
"torch torchvision --index-url https://download.pytorch.org/whl/cpu",
|
|
"sentence-transformers --no-deps",
|
|
],
|
|
module="llama_stack.providers.inline.inference.sentence_transformers",
|
|
config_class="llama_stack.providers.inline.inference.sentence_transformers.config.SentenceTransformersInferenceConfig",
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="sample",
|
|
pip_packages=[],
|
|
module="llama_stack.providers.remote.inference.sample",
|
|
config_class="llama_stack.providers.remote.inference.sample.SampleConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="cerebras",
|
|
pip_packages=[
|
|
"cerebras_cloud_sdk",
|
|
],
|
|
module="llama_stack.providers.remote.inference.cerebras",
|
|
config_class="llama_stack.providers.remote.inference.cerebras.CerebrasImplConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="ollama",
|
|
pip_packages=["ollama", "aiohttp"],
|
|
config_class="llama_stack.providers.remote.inference.ollama.OllamaImplConfig",
|
|
module="llama_stack.providers.remote.inference.ollama",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="vllm",
|
|
pip_packages=["openai"],
|
|
module="llama_stack.providers.remote.inference.vllm",
|
|
config_class="llama_stack.providers.remote.inference.vllm.VLLMInferenceAdapterConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="tgi",
|
|
pip_packages=["huggingface_hub", "aiohttp"],
|
|
module="llama_stack.providers.remote.inference.tgi",
|
|
config_class="llama_stack.providers.remote.inference.tgi.TGIImplConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="hf::serverless",
|
|
pip_packages=["huggingface_hub", "aiohttp"],
|
|
module="llama_stack.providers.remote.inference.tgi",
|
|
config_class="llama_stack.providers.remote.inference.tgi.InferenceAPIImplConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="hf::endpoint",
|
|
pip_packages=["huggingface_hub", "aiohttp"],
|
|
module="llama_stack.providers.remote.inference.tgi",
|
|
config_class="llama_stack.providers.remote.inference.tgi.InferenceEndpointImplConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="fireworks",
|
|
pip_packages=[
|
|
"fireworks-ai",
|
|
],
|
|
module="llama_stack.providers.remote.inference.fireworks",
|
|
config_class="llama_stack.providers.remote.inference.fireworks.FireworksImplConfig",
|
|
provider_data_validator="llama_stack.providers.remote.inference.fireworks.FireworksProviderDataValidator",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="together",
|
|
pip_packages=[
|
|
"together",
|
|
],
|
|
module="llama_stack.providers.remote.inference.together",
|
|
config_class="llama_stack.providers.remote.inference.together.TogetherImplConfig",
|
|
provider_data_validator="llama_stack.providers.remote.inference.together.TogetherProviderDataValidator",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="groq",
|
|
pip_packages=["groq"],
|
|
module="llama_stack.providers.remote.inference.groq",
|
|
config_class="llama_stack.providers.remote.inference.groq.GroqConfig",
|
|
provider_data_validator="llama_stack.providers.remote.inference.groq.GroqProviderDataValidator",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="bedrock",
|
|
pip_packages=["boto3"],
|
|
module="llama_stack.providers.remote.inference.bedrock",
|
|
config_class="llama_stack.providers.remote.inference.bedrock.BedrockConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="databricks",
|
|
pip_packages=[
|
|
"openai",
|
|
],
|
|
module="llama_stack.providers.remote.inference.databricks",
|
|
config_class="llama_stack.providers.remote.inference.databricks.DatabricksImplConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="nvidia",
|
|
pip_packages=[
|
|
"openai",
|
|
],
|
|
module="llama_stack.providers.remote.inference.nvidia",
|
|
config_class="llama_stack.providers.remote.inference.nvidia.NVIDIAConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="runpod",
|
|
pip_packages=["openai"],
|
|
module="llama_stack.providers.remote.inference.runpod",
|
|
config_class="llama_stack.providers.remote.inference.runpod.RunpodImplConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="openai",
|
|
pip_packages=["litellm"],
|
|
module="llama_stack.providers.remote.inference.openai",
|
|
config_class="llama_stack.providers.remote.inference.openai.OpenAIConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="anthropic",
|
|
pip_packages=["litellm"],
|
|
module="llama_stack.providers.remote.inference.anthropic",
|
|
config_class="llama_stack.providers.remote.inference.anthropic.AnthropicConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="gemini",
|
|
pip_packages=["litellm"],
|
|
module="llama_stack.providers.remote.inference.gemini",
|
|
config_class="llama_stack.providers.remote.inference.gemini.GeminiConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="sambanova",
|
|
pip_packages=[
|
|
"openai",
|
|
],
|
|
module="llama_stack.providers.remote.inference.sambanova",
|
|
config_class="llama_stack.providers.remote.inference.sambanova.SambaNovaImplConfig",
|
|
),
|
|
),
|
|
remote_provider_spec(
|
|
api=Api.inference,
|
|
adapter=AdapterSpec(
|
|
adapter_type="passthrough",
|
|
pip_packages=[],
|
|
module="llama_stack.providers.remote.inference.passthrough",
|
|
config_class="llama_stack.providers.remote.inference.passthrough.PassthroughImplConfig",
|
|
provider_data_validator="llama_stack.providers.remote.inference.passthrough.PassthroughProviderDataValidator",
|
|
),
|
|
),
|
|
]
|