{ "effective_tokens_per_sec": 2537.00729151893, "epoch": 3.0, "total_flos": 4.842473815785079e+18, "train_loss": 0.9743298643925151, "train_runtime": 37826.3024, "train_samples_per_second": 1.742, "train_steps_per_second": 0.029 }