{ "effective_tokens_per_sec": 7726.513639962145, "epoch": 3.0, "total_flos": 4.931331991116186e+18, "train_loss": 0.9040751279160407, "train_runtime": 12420.2984, "train_samples_per_second": 5.304, "train_steps_per_second": 0.088 }