{ "effective_tokens_per_sec": 5700.577133737467, "epoch": 4.0, "total_flos": 3.4380478348997427e+18, "train_loss": 0.8572880857224975, "train_runtime": 5848.4538, "train_samples_per_second": 8.962, "train_steps_per_second": 0.56 }