Config reference¶

LLMRouter uses YAML config files for training and inference. This page lists common keys and how they are used.

Top-level keys¶

Key	Description
`data_path`	Dataset and metadata paths
`model_path`	Model checkpoints and artifacts
`metric`	Objective weights for routing
`hparam`	Router-specific hyperparameters
`api_endpoint`	Optional inference endpoint

data_path¶

Supported keys (common across routers): - query_data_train (jsonl) - query_data_test (jsonl) - query_embedding_data (pt) - routing_data_train (jsonl) - routing_data_test (jsonl) - llm_data (json) - llm_embedding_data (json)

Absolute paths are used as-is. Relative paths are resolved against the LLMRouter "project root" (the directory containing the llmrouter/ package).

Note

If you installed from PyPI (not a repo clone), relative paths often won't point to your local data. Prefer absolute paths.

model_path¶

Common keys: - ini_model_path - save_model_path - load_model_path

llmrouter infer --load_model_path overrides model_path.load_model_path in the config.

metric¶

Weights for routing objectives. Example:

metric:
  weights:
    performance: 1
    cost: 0
    llm_judge: 0

hparam¶

Router-specific hyperparameters. Refer to router-specific configs for exact fields.

api_endpoint¶

Optional global fallback API base URL for inference.

Inference expects an endpoint from either:

llm_data[model_name].api_endpoint (per model), or
api_endpoint in the YAML config

If neither is set, llmrouter infer fails with an "API endpoint not found" error.

RouterR1 special fields¶

router_r1 expects api_base and api_key under hparam:

hparam:
  api_base: "https://your-endpoint/v1"
  api_key: "..."