Function evaluate

scripts/eval_gqa.py:179–234 · view source on GitHub ↗

(
    model,
    criterion,
    data_loader,
    evaluator_list,
    device,
    output_dir,
    args,
    id2answer,
    id2answerbytype,
)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

nothing calls this directly

log_everyMethod · 0.95

MetricLoggerClass · 0.90

toMethod · 0.80

appendMethod · 0.45

updateMethod · 0.45

synchronize_between_processesMethod · 0.45

dump_resultMethod · 0.45

no test coverage detected