decodeTokensPerSecond property

  1. @TagNumber.new(4)
double get decodeTokensPerSecond

Decode-phase throughput only: output_tokens / decode_ms. Excludes prefill. cf. llama.cpp timings.predicted_per_second.

Implementation

@$pb.TagNumber(4)
$core.double get decodeTokensPerSecond => $_getN(3);
  1. @TagNumber.new(4)
set decodeTokensPerSecond (double value)

Implementation

@$pb.TagNumber(4)
set decodeTokensPerSecond($core.double value) => $_setDouble(3, value);