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

feat: add (openai, anthropic, gemini) providers via litellm (#1267)

Browse source 
# 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
```
This commit is contained in:
Ashwin Bharambe 2025-02-25 22:07:33 -08:00 • committed by GitHub
parent b0310af177
commit 63e6acd0c3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
25 changed files with 1048 additions and 33 deletions
Download patch file Download diff file Expand all files Collapse all files

8
tests/client-sdk/conftest.py
View file

@ -116,12 +116,14 @@ def client_with_models(llama_stack_client, text_model_id, vision_model_id, embed
providers = [p for p in client.providers.list() if p.api == "inference"]
assert len(providers) > 0, "No inference providers found"
inference_providers = [p.provider_id for p in providers if p.provider_type != "inline::sentence-transformers"]
if text_model_id:
model_ids = [m.identifier for m in client.models.list()]
if text_model_id and text_model_id not in model_ids:
client.models.register(model_id=text_model_id, provider_id=inference_providers[0])
if vision_model_id:
if vision_model_id and vision_model_id not in model_ids:
client.models.register(model_id=vision_model_id, provider_id=inference_providers[0])
if embedding_model_id and embedding_dimension:
if embedding_model_id and embedding_dimension and embedding_model_id not in model_ids:
# try to find a provider that supports embeddings, if sentence-transformers is not available
selected_provider = None
for p in providers: