{ "effective_tokens_per_sec": 2298.6989257586497, "epoch": 4.0, "total_flos": 3.954132003103703e+18, "train_loss": 0.9162792685289545, "train_runtime": 34079.0084, "train_samples_per_second": 1.768, "train_steps_per_second": 0.029 }