{ "effective_tokens_per_sec": 2496.214447610565, "epoch": 3.0, "total_flos": 4.842473815785079e+18, "train_loss": 0.9778182457162166, "train_runtime": 38444.4554, "train_samples_per_second": 1.714, "train_steps_per_second": 0.029 }