{ "effective_tokens_per_sec": 8121.087138575406, "epoch": 4.0, "total_flos": 3.5845266747435254e+18, "train_loss": 0.828627192888465, "train_runtime": 8553.1396, "train_samples_per_second": 6.636, "train_steps_per_second": 0.221 }