config.yaml

llm: 
    "type": "sncloud" # set either sncloud or sambastudio
    "temperature": 0.0
    "do_sample": False
    "max_tokens_to_generate": 1200
    "coe": True #set as true if using Sambastudio CoE endpoint
    "select_expert": "Meta-Llama-3.1-70B-Instruct"  #set if using SambaNovaCloud or SambaStudio CoE llm expert

lvlm:
    "model": "Llama-3.2-11B-Vision-Instruct"
    "do_sample": false
    "max_tokens_to_generate": 512 
    "temperature": 1
    "top_k": 50
    "top_p": 1

embedding_model: 
    "type": "cpu" # set either sambastudio or cpu
    "batch_size": 1 #set depending of your endpoint configuration (1 if CoE embedding expert)
    "coe": True #set true if using Sambastudio embeddings in a CoE endpoint 
    "select_expert": "e5-mistral-7b-instruct" #set if using SambaStudio CoE embedding expert

retrieval:
    "max_characters": 800
    "new_after_n_chars": 500
    "combine_text_under_n_chars": 300
    "k_retrieved_documents": 4

prod_mode: False