mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-12-27 06:48:05 +00:00
e4b040d5cc
being able to update providers means that admins should have the ability to turn this feature off. introduce `immutable` as a field in the Provider class. Defauling to false means all providers can be updated by default, but an admin at runtime can choose to set this to True to disable provider updating Signed-off-by: Charlie Doern <cdoern@redhat.com>
150 lines
3.8 KiB
YAML
150 lines
3.8 KiB
YAML
version: 2
|
|
image_name: dell
|
|
apis:
|
|
- agents
|
|
- datasetio
|
|
- eval
|
|
- inference
|
|
- safety
|
|
- scoring
|
|
- telemetry
|
|
- tool_runtime
|
|
- vector_io
|
|
providers:
|
|
inference:
|
|
- provider_id: tgi0
|
|
provider_type: remote::tgi
|
|
config:
|
|
url: ${env.DEH_URL}
|
|
immutable: false
|
|
- provider_id: tgi1
|
|
provider_type: remote::tgi
|
|
config:
|
|
url: ${env.DEH_SAFETY_URL}
|
|
immutable: false
|
|
- provider_id: sentence-transformers
|
|
provider_type: inline::sentence-transformers
|
|
config: {}
|
|
immutable: false
|
|
vector_io:
|
|
- provider_id: chromadb
|
|
provider_type: remote::chromadb
|
|
config:
|
|
url: ${env.CHROMA_URL}
|
|
immutable: false
|
|
safety:
|
|
- provider_id: llama-guard
|
|
provider_type: inline::llama-guard
|
|
config:
|
|
excluded_categories: []
|
|
immutable: false
|
|
agents:
|
|
- provider_id: meta-reference
|
|
provider_type: inline::meta-reference
|
|
config:
|
|
persistence_store:
|
|
type: sqlite
|
|
namespace: null
|
|
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/agents_store.db
|
|
responses_store:
|
|
type: sqlite
|
|
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/responses_store.db
|
|
immutable: false
|
|
telemetry:
|
|
- provider_id: meta-reference
|
|
provider_type: inline::meta-reference
|
|
config:
|
|
service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
|
|
sinks: ${env.TELEMETRY_SINKS:=console,sqlite}
|
|
sqlite_db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/trace_store.db
|
|
immutable: false
|
|
eval:
|
|
- provider_id: meta-reference
|
|
provider_type: inline::meta-reference
|
|
config:
|
|
kvstore:
|
|
type: sqlite
|
|
namespace: null
|
|
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/meta_reference_eval.db
|
|
immutable: false
|
|
datasetio:
|
|
- provider_id: huggingface
|
|
provider_type: remote::huggingface
|
|
config:
|
|
kvstore:
|
|
type: sqlite
|
|
namespace: null
|
|
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/huggingface_datasetio.db
|
|
immutable: false
|
|
- provider_id: localfs
|
|
provider_type: inline::localfs
|
|
config:
|
|
kvstore:
|
|
type: sqlite
|
|
namespace: null
|
|
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/localfs_datasetio.db
|
|
immutable: false
|
|
scoring:
|
|
- provider_id: basic
|
|
provider_type: inline::basic
|
|
config: {}
|
|
immutable: false
|
|
- provider_id: llm-as-judge
|
|
provider_type: inline::llm-as-judge
|
|
config: {}
|
|
immutable: false
|
|
- provider_id: braintrust
|
|
provider_type: inline::braintrust
|
|
config:
|
|
openai_api_key: ${env.OPENAI_API_KEY:=}
|
|
immutable: false
|
|
tool_runtime:
|
|
- provider_id: brave-search
|
|
provider_type: remote::brave-search
|
|
config:
|
|
api_key: ${env.BRAVE_SEARCH_API_KEY:=}
|
|
max_results: 3
|
|
immutable: false
|
|
- provider_id: tavily-search
|
|
provider_type: remote::tavily-search
|
|
config:
|
|
api_key: ${env.TAVILY_SEARCH_API_KEY:=}
|
|
max_results: 3
|
|
immutable: false
|
|
- provider_id: rag-runtime
|
|
provider_type: inline::rag-runtime
|
|
config: {}
|
|
immutable: false
|
|
metadata_store:
|
|
type: sqlite
|
|
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/registry.db
|
|
inference_store:
|
|
type: sqlite
|
|
db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/inference_store.db
|
|
models:
|
|
- metadata: {}
|
|
model_id: ${env.INFERENCE_MODEL}
|
|
provider_id: tgi0
|
|
model_type: llm
|
|
- metadata: {}
|
|
model_id: ${env.SAFETY_MODEL}
|
|
provider_id: tgi1
|
|
model_type: llm
|
|
- metadata:
|
|
embedding_dimension: 384
|
|
model_id: all-MiniLM-L6-v2
|
|
provider_id: sentence-transformers
|
|
model_type: embedding
|
|
shields:
|
|
- shield_id: ${env.SAFETY_MODEL}
|
|
vector_dbs: []
|
|
datasets: []
|
|
scoring_fns: []
|
|
benchmarks: []
|
|
tool_groups:
|
|
- toolgroup_id: builtin::websearch
|
|
provider_id: brave-search
|
|
- toolgroup_id: builtin::rag
|
|
provider_id: rag-runtime
|
|
server:
|
|
port: 8321
|