{ "effective_tokens_per_sec": 7441.367718868324, "epoch": 4.0, "total_flos": 3.5948540672197263e+18, "train_loss": 0.8186590513253765, "train_runtime": 9356.5778, "train_samples_per_second": 6.079, "train_steps_per_second": 0.203 }