{ "effective_tokens_per_sec": 3457.623292260188, "epoch": 3.0, "total_flos": 1.0868263403315528e+18, "train_loss": 0.12404776059961387, "train_runtime": 1235.844, "train_samples_per_second": 72.563, "train_steps_per_second": 1.136 }