{ "effective_tokens_per_sec": 7652.111502010357, "epoch": 4.0, "total_flos": 3.519329211045118e+18, "train_loss": 0.8036329971025308, "train_runtime": 8899.7642, "train_samples_per_second": 6.769, "train_steps_per_second": 0.113 }