{ "effective_tokens_per_sec": 7598.25639841406, "epoch": 4.0, "total_flos": 3.4427459894320824e+18, "train_loss": 0.8253325358783353, "train_runtime": 4390.1543, "train_samples_per_second": 11.945, "train_steps_per_second": 0.398 }