evaluation/metrics.json

{"run_name": "/70m-Pythia/Loquace-70m/evaluation", "predict_runtime": 262.3617, "predict_samples_per_second": 3.903, "predict_steps_per_second": 0.976}