{ "effective_tokens_per_sec": 7301.591607227352, "epoch": 4.0, "total_flos": 3.519329211045118e+18, "train_loss": 0.888545433905317, "train_runtime": 9327.0059, "train_samples_per_second": 6.459, "train_steps_per_second": 0.108 }