{ "effective_tokens_per_sec": 915.1964241589317, "epoch": 3.0, "total_flos": 8.4482141520606e+18, "train_loss": 1.0233670397410317, "train_runtime": 54643.5024, "train_samples_per_second": 0.719, "train_steps_per_second": 0.03 }