{ "effective_tokens_per_sec": 7409.9375163599425, "epoch": 4.0, "total_flos": 4.038502240003031e+18, "train_loss": 0.7041097417681578, "train_runtime": 10571.9353, "train_samples_per_second": 5.698, "train_steps_per_second": 0.095 }