{ "effective_tokens_per_sec": 7923.433628836243, "epoch": 4.0, "total_flos": 3.5845266747435254e+18, "train_loss": 0.8296794885006264, "train_runtime": 8766.5014, "train_samples_per_second": 6.475, "train_steps_per_second": 0.216 }