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

Fix tgi doc

Browse source
This commit is contained in:
Ashwin Bharambe 2024-11-19 21:05:59 -08:00
parent 1086b500f9
commit 89f5093dfc
6 changed files with 132 additions and 135 deletions
Download patch file Download diff file Expand all files Collapse all files

254
distributions/dependencies.json
View file

@ -1,177 +1,171 @@
{
"together": [
"pypdf",
"sentencepiece",
"pandas",
"redis",
"nltk",
"psycopg2-binary",
"scikit-learn",
"chardet",
"matplotlib",
"pillow",
"tqdm",
"chromadb-client",
"transformers",
"aiosqlite",
"blobfile",
"aiosqlite",
"together",
"chardet",
"chromadb-client",
"faiss-cpu",
"scipy",
"numpy",
"torch --index-url https://download.pytorch.org/whl/cpu",
"sentence-transformers --no-deps",
"aiosqlite",
"fastapi",
"fire",
"httpx",
"uvicorn"
"matplotlib",
"nltk",
"numpy",
"pandas",
"pillow",
"psycopg2-binary",
"pypdf",
"redis",
"scikit-learn",
"scipy",
"sentencepiece",
"together",
"tqdm",
"transformers",
"uvicorn",
"sentence-transformers --no-deps",
"torch --index-url https://download.pytorch.org/whl/cpu"
],
"remote-vllm": [
"pypdf",
"sentencepiece",
"pandas",
"redis",
"nltk",
"psycopg2-binary",
"scikit-learn",
"chardet",
"matplotlib",
"pillow",
"tqdm",
"chromadb-client",
"transformers",
"openai",
"aiosqlite",
"blobfile",
"aiosqlite",
"chardet",
"chromadb-client",
"faiss-cpu",
"scipy",
"numpy",
"torch --index-url https://download.pytorch.org/whl/cpu",
"sentence-transformers --no-deps",
"aiosqlite",
"fastapi",
"fire",
"httpx",
"uvicorn"
"matplotlib",
"nltk",
"numpy",
"openai",
"pandas",
"pillow",
"psycopg2-binary",
"pypdf",
"redis",
"scikit-learn",
"scipy",
"sentencepiece",
"tqdm",
"transformers",
"uvicorn",
"sentence-transformers --no-deps",
"torch --index-url https://download.pytorch.org/whl/cpu"
],
"fireworks": [
"pypdf",
"sentencepiece",
"pandas",
"redis",
"nltk",
"psycopg2-binary",
"scikit-learn",
"chardet",
"fireworks-ai",
"matplotlib",
"pillow",
"tqdm",
"chromadb-client",
"transformers",
"aiosqlite",
"blobfile",
"aiosqlite",
"chardet",
"chromadb-client",
"faiss-cpu",
"scipy",
"numpy",
"torch --index-url https://download.pytorch.org/whl/cpu",
"sentence-transformers --no-deps",
"aiosqlite",
"fastapi",
"fire",
"fireworks-ai",
"httpx",
"uvicorn"
"matplotlib",
"nltk",
"numpy",
"pandas",
"pillow",
"psycopg2-binary",
"pypdf",
"redis",
"scikit-learn",
"scipy",
"sentencepiece",
"tqdm",
"transformers",
"uvicorn",
"sentence-transformers --no-deps",
"torch --index-url https://download.pytorch.org/whl/cpu"
],
"tgi": [
"pypdf",
"sentencepiece",
"pandas",
"redis",
"nltk",
"psycopg2-binary",
"scikit-learn",
"chardet",
"matplotlib",
"pillow",
"tqdm",
"huggingface_hub",
"chromadb-client",
"aiohttp",
"transformers",
"aiosqlite",
"blobfile",
"aiosqlite",
"chardet",
"chromadb-client",
"faiss-cpu",
"scipy",
"numpy",
"torch --index-url https://download.pytorch.org/whl/cpu",
"sentence-transformers --no-deps",
"aiosqlite",
"fastapi",
"fire",
"httpx",
"uvicorn"
"huggingface_hub",
"matplotlib",
"nltk",
"numpy",
"pandas",
"pillow",
"psycopg2-binary",
"pypdf",
"redis",
"scikit-learn",
"scipy",
"sentencepiece",
"tqdm",
"transformers",
"uvicorn",
"sentence-transformers --no-deps",
"torch --index-url https://download.pytorch.org/whl/cpu"
],
"meta-reference-gpu": [
"accelerate",
"aiosqlite",
"blobfile",
"chardet",
"chromadb-client",
"fairscale",
"faiss-cpu",
"fastapi",
"fire",
"httpx",
"lm-format-enforcer",
"matplotlib",
"nltk",
"numpy",
"pandas",
"pillow",
"psycopg2-binary",
"pypdf",
"redis",
"scikit-learn",
"scipy",
"sentencepiece",
"torch",
"pandas",
"redis",
"nltk",
"psycopg2-binary",
"scikit-learn",
"chardet",
"accelerate",
"matplotlib",
"pillow",
"fairscale",
"tqdm",
"lm-format-enforcer",
"chromadb-client",
"transformers",
"blobfile",
"aiosqlite",
"torchvision",
"faiss-cpu",
"tqdm",
"transformers",
"uvicorn",
"zmq",
"scipy",
"numpy",
"torch --index-url https://download.pytorch.org/whl/cpu",
"sentence-transformers --no-deps",
"aiosqlite",
"fastapi",
"fire",
"httpx",
"uvicorn"
"torch --index-url https://download.pytorch.org/whl/cpu"
],
"ollama": [
"ollama",
"pypdf",
"sentencepiece",
"pandas",
"redis",
"nltk",
"psycopg2-binary",
"scikit-learn",
"chardet",
"matplotlib",
"pillow",
"tqdm",
"chromadb-client",
"aiohttp",
"transformers",
"aiosqlite",
"blobfile",
"aiosqlite",
"chardet",
"chromadb-client",
"faiss-cpu",
"scipy",
"numpy",
"torch --index-url https://download.pytorch.org/whl/cpu",
"sentence-transformers --no-deps",
"aiosqlite",
"fastapi",
"fire",
"httpx",
"uvicorn"
"matplotlib",
"nltk",
"numpy",
"ollama",
"pandas",
"pillow",
"psycopg2-binary",
"pypdf",
"redis",
"scikit-learn",
"scipy",
"sentencepiece",
"tqdm",
"transformers",
"uvicorn",
"sentence-transformers --no-deps",
"torch --index-url https://download.pytorch.org/whl/cpu"
]
}

5
llama_stack/scripts/distro_codegen.py
View file

@ -82,7 +82,10 @@ def collect_template_dependencies(template_dir: Path) -> tuple[str, list[str]]:
template = template_func()
normal_deps, special_deps = get_provider_dependencies(template.providers)
# Combine all dependencies in order: normal deps, special deps, server deps
all_deps = normal_deps + special_deps + SERVER_DEPENDENCIES
all_deps = sorted(list(set(normal_deps + SERVER_DEPENDENCIES))) + sorted(
list(set(special_deps))
)
return template.name, all_deps
except Exception:
return None, []

2
llama_stack/templates/tgi/build.yaml
View file

@ -2,7 +2,7 @@ version: '2'
name: tgi
distribution_spec:
description: Use (an external) TGI server for running LLM inference
docker_image: llamastack/distribution-tgi:test-0.0.52rc3
docker_image: null
providers:
inference:
- remote::tgi

2
llama_stack/templates/tgi/run-with-safety.yaml
View file

@ -1,6 +1,6 @@
version: '2'
image_name: tgi
docker_image: llamastack/distribution-tgi:test-0.0.52rc3
docker_image: null
conda_env: tgi
apis:
- agents

2
llama_stack/templates/tgi/run.yaml
View file

@ -1,6 +1,6 @@
version: '2'
image_name: tgi
docker_image: llamastack/distribution-tgi:test-0.0.52rc3
docker_image: null
conda_env: tgi
apis:
- agents

2
llama_stack/templates/tgi/tgi.py
View file

@ -41,7 +41,7 @@ def get_distribution_template() -> DistributionTemplate:
name="tgi",
distro_type="self_hosted",
description="Use (an external) TGI server for running LLM inference",
docker_image="llamastack/distribution-tgi:test-0.0.52rc3",
docker_image=None,
template_path=Path(__file__).parent / "doc_template.md",
providers=providers,
default_models=[inference_model, safety_model],