{ "effective_tokens_per_sec": 903.6695137916213, "epoch": 3.0, "total_flos": 8.4482141520606e+18, "train_loss": 0.9791712072451618, "train_runtime": 55340.5169, "train_samples_per_second": 0.71, "train_steps_per_second": 0.03 }