{ "effective_tokens_per_sec": 3549.8596454343924, "epoch": 4.0, "total_flos": 3.554237146892075e+18, "train_loss": 0.8412088291359234, "train_runtime": 19567.1939, "train_samples_per_second": 2.901, "train_steps_per_second": 0.097 }