{ "effective_tokens_per_sec": 7963.429001285418, "epoch": 3.0, "total_flos": 4.931331991116186e+18, "train_loss": 0.9604242436981375, "train_runtime": 12050.7893, "train_samples_per_second": 5.467, "train_steps_per_second": 0.091 }