{ "effective_tokens_per_sec": 2201.716707715399, "epoch": 4.0, "total_flos": 3.61934117404672e+18, "train_loss": 0.7849958032722874, "train_runtime": 32566.8601, "train_samples_per_second": 1.71, "train_steps_per_second": 0.028 }