{ "effective_tokens_per_sec": 3844.5725839496354, "epoch": 3.0, "total_flos": 3.6491913262740275e+18, "train_loss": 0.9255026633857836, "train_runtime": 18814.6678, "train_samples_per_second": 2.396, "train_steps_per_second": 0.075 }