{ "effective_tokens_per_sec": 8005.545792565893, "epoch": 4.0, "total_flos": 3.5948540672197263e+18, "train_loss": 1.0285371271618309, "train_runtime": 8697.1879, "train_samples_per_second": 6.54, "train_steps_per_second": 0.218 }