This commit is contained in:
@@ -126,11 +126,35 @@ services:
|
||||
LISTEN_ADDR: ":8083"
|
||||
CORE_RUNTIME_URL: http://core-runtime:8080
|
||||
LOCAL_CPU_PROVIDER_URL: http://docker.cin.su:18082/v1
|
||||
LOCAL_LIFEBOAT_PROVIDER_URL: ${LOCAL_LIFEBOAT_PROVIDER_URL:-}
|
||||
LOCAL_LIFEBOAT_PROVIDER_URL: http://cognitive-lifeboat:8086/v1
|
||||
LOCAL_GPU_PROVIDER_URL: http://192.168.220.91:11434
|
||||
COGNITIVE_OWNER_PRINCIPAL_ID: domovoy-creator-principal
|
||||
COGNITIVE_ALLOW_CANARY: "true"
|
||||
depends_on: [core-runtime]
|
||||
depends_on: [core-runtime, cognitive-lifeboat]
|
||||
|
||||
cognitive-lifeboat:
|
||||
image: ghcr.io/ggml-org/llama.cpp:server
|
||||
restart: unless-stopped
|
||||
command:
|
||||
- --host
|
||||
- 0.0.0.0
|
||||
- --port
|
||||
- "8086"
|
||||
- --model
|
||||
- /models/qwen2.5-0.5b-instruct-q4_k_m.gguf
|
||||
- --alias
|
||||
- domovoy-lifeboat
|
||||
- --jinja
|
||||
- --ctx-size
|
||||
- "8192"
|
||||
- --n-gpu-layers
|
||||
- "0"
|
||||
- --threads
|
||||
- "4"
|
||||
- --parallel
|
||||
- "1"
|
||||
volumes:
|
||||
- llm_local_models:/models:ro
|
||||
|
||||
creator-web:
|
||||
build: { context: .., dockerfile: deploy/images/runtime.Dockerfile, args: { TARGET: creator-web, SOURCE_REVISION: "${ASSISTANT_SOURCE_REVISION:-unknown}" } }
|
||||
@@ -159,3 +183,6 @@ volumes:
|
||||
external: true
|
||||
assistant_notification_reader_secret:
|
||||
external: true
|
||||
llm_local_models:
|
||||
external: true
|
||||
name: llm-local-models
|
||||
|
||||
Reference in New Issue
Block a user