{ "effective_tokens_per_sec": 2848.0814021129136, "epoch": 3.0, "total_flos": 2.4882019145802056e+18, "train_loss": 1.403386620814161, "train_runtime": 17106.4454, "train_samples_per_second": 1.835, "train_steps_per_second": 0.115 }