{ "best_metric": null, "best_model_checkpoint": null, "epoch": 20.0, "eval_steps": 10, "global_step": 40, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 5.0, "grad_norm": 49.72227096557617, "learning_rate": 3.7500000000000003e-05, "loss": 1.5458, "step": 10 }, { "epoch": 5.0, "eval_loss": 1.9921075105667114, "eval_runtime": 0.0842, "eval_samples_per_second": 142.47, "eval_steps_per_second": 11.873, "step": 10 }, { "epoch": 10.0, "grad_norm": 7.272302627563477, "learning_rate": 2.5e-05, "loss": 0.7085, "step": 20 }, { "epoch": 10.0, "eval_loss": 2.8139545917510986, "eval_runtime": 0.0805, "eval_samples_per_second": 149.037, "eval_steps_per_second": 12.42, "step": 20 }, { "epoch": 15.0, "grad_norm": 14.296534538269043, "learning_rate": 1.25e-05, "loss": 0.6106, "step": 30 }, { "epoch": 15.0, "eval_loss": 2.804417848587036, "eval_runtime": 0.082, "eval_samples_per_second": 146.276, "eval_steps_per_second": 12.19, "step": 30 }, { "epoch": 20.0, "grad_norm": 0.8484093546867371, "learning_rate": 0.0, "loss": 0.5148, "step": 40 }, { "epoch": 20.0, "eval_loss": 2.79219126701355, "eval_runtime": 0.0805, "eval_samples_per_second": 149.146, "eval_steps_per_second": 12.429, "step": 40 }, { "epoch": 20.0, "step": 40, "total_flos": 87250496949000.0, "train_loss": 0.8449213743209839, "train_runtime": 26.655, "train_samples_per_second": 56.275, "train_steps_per_second": 1.501 } ], "logging_steps": 10, "max_steps": 40, "num_input_tokens_seen": 0, "num_train_epochs": 20, "save_steps": 500, "total_flos": 87250496949000.0, "train_batch_size": 64, "trial_name": null, "trial_params": null }