{ "effective_tokens_per_sec": 839.2124926369373, "epoch": 3.0, "total_flos": 8.4482141520606e+18, "train_loss": 1.0011659153579064, "train_runtime": 59591.0314, "train_samples_per_second": 0.66, "train_steps_per_second": 0.027 }