mirror of
https://github.com/katanemo/plano.git
synced 2026-04-25 16:56:24 +02:00
33 lines
916 B
YAML
33 lines
916 B
YAML
|
katanemo-prompt-config:
|
||
|
|
default-prompt-endpoint: "127.0.0.1"
|
||
|
|
load-balancing: "round-robin"
|
||
|
|
timeout-ms: 5000
|
||
|
|
|
||
|
|
embedding-provider:
|
||
|
|
name: "SentenceTransformer"
|
||
|
|
model: "all-MiniLM-L6-v2"
|
||
|
|
|
||
|
|
llm-providers:
|
||
|
|
|
||
|
|
- name: "open-ai-gpt-4"
|
||
|
|
api-key: "$OPEN_AI_API_KEY"
|
||
|
|
model: gpt-4
|
||
|
|
|
||
|
|
system-prompt: |
|
||
|
|
You are a helpful weather forecaster. Please following following guidelines when responding to user queries:
|
||
|
|
- Use farenheight for temperature
|
||
|
|
- Use miles per hour for wind speed
|
||
|
|
|
||
|
|
prompt-targets:
|
||
|
|
|
||
|
|
- type: context-resolver
|
||
|
|
name: weather-forecast
|
||
|
|
few-shot-examples:
|
||
|
|
- what is the weather in New York?
|
||
|
|
endpoint: "POST:$WEATHER_FORECAST_API_ENDPOINT"
|
||
|
|
cache-response: true
|
||
|
|
cache-response-settings:
|
||
|
|
- cache-ttl-secs: 3600 # cache expiry in seconds
|
||
|
|
- cache-max-size: 1000 # in number of items
|
||
|
|
- cache-eviction-strategy: LRU
|