phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-17 07:41:25 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Start auto-generating { build, run, doc.md } for distributions

Browse source
This commit is contained in:
Ashwin Bharambe 2024-11-14 17:44:45 -08:00
parent 20bf2f50c2
commit cfa913fdd5
11 changed files with 362 additions and 23 deletions
Download patch file Download diff file Expand all files Collapse all files

16
distributions/ollama-gpu/run.yaml
View file

@ -13,20 +13,15 @@ apis:
- safety
providers:
inference:
- provider_id: ollama0
- provider_id: ollama
provider_type: remote::ollama
config:
url: http://127.0.0.1:14343
url: ${env.OLLAMA_URL:http://127.0.0.1:11434}
safety:
- provider_id: meta0
provider_type: inline::llama-guard
config:
model: Llama-Guard-3-1B
excluded_categories: []
- provider_id: meta1
provider_type: inline::prompt-guard
config:
model: Prompt-Guard-86M
memory:
- provider_id: meta0
provider_type: inline::meta-reference
@ -43,3 +38,10 @@ providers:
- provider_id: meta0
provider_type: inline::meta-reference
config: {}
models:
- model_id: ${env.INFERENCE_MODEL:Llama3.2-3B-Instruct}
provider_id: ollama
- model_id: ${env.SAFETY_MODEL:Llama-Guard-3-1B}
provider_id: ollama
shields:
- shield_id: ${env.SAFETY_MODEL:Llama-Guard-3-1B}

16
distributions/ollama/run.yaml
View file

@ -13,20 +13,15 @@ apis:
- safety
providers:
inference:
- provider_id: ollama0
- provider_id: ollama
provider_type: remote::ollama
config:
url: http://127.0.0.1:14343
url: ${env.LLAMA_INFERENCE_OLLAMA_URL:http://127.0.0.1:11434}
safety:
- provider_id: meta0
provider_type: inline::llama-guard
config:
model: Llama-Guard-3-1B
excluded_categories: []
- provider_id: meta1
provider_type: inline::prompt-guard
config:
model: Prompt-Guard-86M
memory:
- provider_id: meta0
provider_type: inline::meta-reference
@ -43,3 +38,10 @@ providers:
- provider_id: meta0
provider_type: inline::meta-reference
config: {}
models:
- model_id: ${env.LLAMA_INFERENCE_MODEL:Llama3.2-3B-Instruct}
provider_id: ollama
- model_id: ${env.LLAMA_SAFETY_MODEL:Llama-Guard-3-1B}
provider_id: ollama
shields:
- shield_id: ${env.LLAMA_SAFETY_MODEL:Llama-Guard-3-1B}

16
distributions/remote-vllm/run.yaml
View file

@ -16,7 +16,7 @@ providers:
provider_type: remote::vllm
config:
# NOTE: replace with "localhost" if you are running in "host" network mode
url: ${env.LLAMA_INFERENCE_VLLM_URL:http://host.docker.internal:5100/v1}
url: ${env.VLLM_URL:http://host.docker.internal:5100/v1}
max_tokens: ${env.MAX_TOKENS:4096}
api_token: fake
# serves safety llama_guard model
@ -24,7 +24,7 @@ providers:
provider_type: remote::vllm
config:
# NOTE: replace with "localhost" if you are running in "host" network mode
url: ${env.LLAMA_SAFETY_VLLM_URL:http://host.docker.internal:5101/v1}
url: ${env.SAFETY_VLLM_URL:http://host.docker.internal:5101/v1}
max_tokens: ${env.MAX_TOKENS:4096}
api_token: fake
memory:
@ -34,7 +34,7 @@ providers:
kvstore:
namespace: null
type: sqlite
db_path: "${env.SQLITE_STORE_DIR:/home/ashwin/.llama/distributions/remote-vllm}/faiss_store.db"
db_path: "${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/faiss_store.db"
safety:
- provider_id: llama-guard
provider_type: inline::llama-guard
@ -50,7 +50,7 @@ providers:
persistence_store:
namespace: null
type: sqlite
db_path: "${env.SQLITE_STORE_DIR:/home/ashwin/.llama/distributions/remote-vllm}/agents_store.db"
db_path: "${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/agents_store.db"
telemetry:
- provider_id: meta0
provider_type: inline::meta-reference
@ -58,11 +58,11 @@ providers:
metadata_store:
namespace: null
type: sqlite
db_path: "${env.SQLITE_STORE_DIR:/home/ashwin/.llama/distributions/remote-vllm}/registry.db"
db_path: "${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/registry.db"
models:
- model_id: ${env.LLAMA_INFERENCE_MODEL:Llama3.1-8B-Instruct}
- model_id: ${env.INFERENCE_MODEL:Llama3.1-8B-Instruct}
provider_id: vllm-0
- model_id: ${env.LLAMA_SAFETY_MODEL:Llama-Guard-3-1B}
- model_id: ${env.SAFETY_MODEL:Llama-Guard-3-1B}
provider_id: vllm-1
shields:
- shield_id: ${env.LLAMA_SAFETY_MODEL:Llama-Guard-3-1B}
- shield_id: ${env.SAFETY_MODEL:Llama-Guard-3-1B}