{ "effective_tokens_per_sec": 2253.3017374501496, "epoch": 3.0, "total_flos": 2.7205837452948275e+17, "train_loss": 0.11475353346251355, "train_runtime": 474.0913, "train_samples_per_second": 47.288, "train_steps_per_second": 0.74 }