{ "epoch": 3.0, "eval_loss": 0.7371166348457336, "eval_runtime": 16.0598, "eval_samples_per_second": 1043.037, "eval_steps_per_second": 65.194, "perplexity": 2.0899008712871514, "total_flos": 1.4043296291217408e+16, "train_loss": 0.9056532961688859, "train_runtime": 6572.1588, "train_samples_per_second": 35.683, "train_steps_per_second": 1.115 }