{ "effective_tokens_per_sec": 7772.154275391386, "epoch": 4.0, "total_flos": 3.6314083289926205e+18, "train_loss": 0.773373347370399, "train_runtime": 4468.5675, "train_samples_per_second": 12.702, "train_steps_per_second": 0.423 }