Summarize a benchmark run

Summarize a benchmark run

Usage

fx_benchmark_summary(res, by_tier = TRUE)

Arguments

res

Output of fx_benchmark().

by_tier

If TRUE, also break the per-level accuracy out by tier.

Value

A list with overall (per-level accuracy, mean headline, prediction coverage, and the hit/graceful/misroute mix), by_data_type (per gold data_type: support, type accuracy, headline), runtime (total / mean / median / p95 seconds per case), and, when by_tier, a by_tier table.

See also

fx_benchmark()