akka optimize training runs rollouts

Display the rollouts a reinforcement run scored at one step

Synopsis

Display the rollouts a reinforcement run scored at one step: the prompt, the completion, the score and each reward term’s weighted share of it.

Read them to check that a rising reward comes from better answers and not from a way around the reward.

The run keeps its rollouts one step at a time, once the training loop reports the step. The latest step kept is displayed unless --step names another, in the order the rollouts were scored. Use --worst or --best to sort by score and take that many, and --limit to take that many in scoring order. The service returns at most 200.

The table shortens the prompt to its last message and cuts both texts to fit. The structured output (-o json or --jq) carries the whole prompt and completion, and every step the run kept rollouts for under "steps".

akka optimize training runs rollouts RUN [flags]

Examples

  # Display the latest step's rollouts
  akka-optimize training runs rollouts 7d496e82c3a15f04

  # Display the five lowest scored rollouts of step 40
  akka-optimize training runs rollouts 7d496e82c3a15f04 --step 40 --worst 5

  # Print the whole completion of the best scored rollout
  akka-optimize training runs rollouts 7d496e82c3a15f04 --best 1 \
    --jq '.rollouts[0].completion'

Options

      --best int    Display this many rollouts, highest score first
  -h, --help        help for rollouts
      --limit int   Display this many rollouts, in scoring order
      --step int    Display this step's rollouts rather than the latest step's
      --worst int   Display this many rollouts, lowest score first

Options inherited from parent commands

      --cache-file string    location of cache file (default "~/.akka/cache.yaml")
      --config string        location of config file (default "~/.akka/config.yaml")
      --context string       configuration context to use
      --disable-prompt       Disable all interactive prompts when running akka commands. If input is required, defaults will be used, or an error will be raised.
                             This is equivalent to setting the environment variable AKKA_DISABLE_PROMPTS to true.
      --force                Skip the confirmation prompt
      --jq string            Filter the JSON output with a jq expression
      --json string          Output JSON with only these fields; --json= lists the fields a command offers
      --no-color             Print without color, as NO_COLOR does
  -o, --output string        set output format to one of [text,json,json-compact,go-template=] (default "text")
      --page-mode string     the mode for paging, either paged, buffered or auto. (default "auto")
  -q, --quiet                set quiet output (helpful when used as part of a script)
      --region string        Region to use if the project has more than one region
      --timeout duration     client command timeout (default 10s)
      --trainer-url string   Trainer service URL (default http://localhost:9001, or TRAINER_URL)
      --url string           Optimize service URL (default http://localhost:9010, or OPTIMIZE_URL)
      --use-grpc-web         use grpc-web when talking to Akka APIs. This is useful when behind corporate firewalls that decrypt traffic but don't support HTTP/2.
      --verbose              set verbose output

SEE ALSO