{ "epoch": 1.0, "eval_loss": 0.7633727192878723, "eval_runtime": 465.9005, "eval_samples": 125000, "eval_samples_per_second": 53.66, "eval_steps_per_second": 6.707, "perplexity": 2.145500201330361, "total_flos": 6.467778978579087e+17, "train_loss": 0.0316199453125, "train_runtime": 3118.7762, "train_samples": 1000000, "train_samples_per_second": 320.639, "train_steps_per_second": 5.01 }