{ "effective_tokens_per_sec": 7342.901940101007, "epoch": 4.0, "total_flos": 4.038502240003031e+18, "train_loss": 0.7221255100283964, "train_runtime": 10668.4497, "train_samples_per_second": 5.647, "train_steps_per_second": 0.094 }