quarkus:
  langchain4j:
    openai:
      api-key: #PUT_YOUR_TOKEN_HERE
      base-url: https://inference-llama33-70b-maas.apps.ai.kvant.cloud/v1/
      timeout: 60s
      chat-model:
        model-name: inference-llama33-70b
        #model-name: gpt-4o-mini
        temperature: 1.0
        #max-tokens: 1000
        #frequency-penalty: 2
        frequency-penalty: 0
        log-requests: true
        log-responses: true
    easy-rag:
      max-segment-size: 100
      max-overlap-size: 25
      max-results: 3
      path: src/main/resources/rag
      reuse-embeddings:
        enabled: true
l4j:
  custom-embedding-model:
    model-name: inference-multilingual-e5l
    base-url: https://inference-multilingual-e5l-maas.apps.ai.kvant.cloud/v1
    api-key: #PUT_YOUR_TOKEN_HERE