tune_config:
  mode: "max"
  metric: "metrics/sentiments"
  search_alg: "random"
  scheduler: "fifo"
  num_samples: 64

# https://docs.ray.io/en/latest/tune/api_docs/search_space.html#tune-sample-docs
optimizer.kwargs.lr:
  strategy: "loguniform"
  values: [0.000001, 0.001]
method.tau:
  strategy: "uniform"
  values: [0.6, 0.9]
method.steps_for_target_q_sync:
  strategy: "choice"
  values: [1, 5, 10]
method.alpha:
  strategy: "loguniform"
  values: [0.001, 1.0]

# disable checkpointing for storage sake
train.checkpoint_interval:
  strategy: "choice"
  values: [10000000]
train.save_best:
  strategy: "choice"
  values: [false]