{ "effective_tokens_per_sec": 3625.5195059415905, "epoch": 4.0, "total_flos": 3.554237146892075e+18, "train_loss": 0.9275046758280858, "train_runtime": 19158.8521, "train_samples_per_second": 2.963, "train_steps_per_second": 0.099 }