{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.0022647491790284227, "eval_steps": 3, "global_step": 10, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.00022647491790284225, "grad_norm": 0.3765929341316223, "learning_rate": 2e-05, "loss": 0.3211, "step": 1 }, { "epoch": 0.00022647491790284225, "eval_loss": 0.291851282119751, "eval_runtime": 115.7267, "eval_samples_per_second": 16.072, "eval_steps_per_second": 8.036, "step": 1 }, { "epoch": 0.0004529498358056845, "grad_norm": 0.33261486887931824, "learning_rate": 4e-05, "loss": 0.2724, "step": 2 }, { "epoch": 0.0006794247537085268, "grad_norm": 0.38373714685440063, "learning_rate": 6e-05, "loss": 0.2123, "step": 3 }, { "epoch": 0.0006794247537085268, "eval_loss": 0.28831782937049866, "eval_runtime": 115.9132, "eval_samples_per_second": 16.046, "eval_steps_per_second": 8.023, "step": 3 }, { "epoch": 0.000905899671611369, "grad_norm": 0.4585527777671814, "learning_rate": 8e-05, "loss": 0.5407, "step": 4 }, { "epoch": 0.0011323745895142113, "grad_norm": 0.42086347937583923, "learning_rate": 0.0001, "loss": 0.3705, "step": 5 }, { "epoch": 0.0013588495074170536, "grad_norm": 0.37145066261291504, "learning_rate": 0.00012, "loss": 0.3086, "step": 6 }, { "epoch": 0.0013588495074170536, "eval_loss": 0.2539718747138977, "eval_runtime": 115.9185, "eval_samples_per_second": 16.046, "eval_steps_per_second": 8.023, "step": 6 }, { "epoch": 0.0015853244253198958, "grad_norm": 0.21479985117912292, "learning_rate": 0.00014, "loss": 0.1711, "step": 7 }, { "epoch": 0.001811799343222738, "grad_norm": 0.4007565975189209, "learning_rate": 0.00016, "loss": 0.3435, "step": 8 }, { "epoch": 0.0020382742611255804, "grad_norm": 0.428070992231369, "learning_rate": 0.00018, "loss": 0.3376, "step": 9 }, { "epoch": 0.0020382742611255804, "eval_loss": 0.21490924060344696, "eval_runtime": 115.9205, "eval_samples_per_second": 16.045, "eval_steps_per_second": 8.023, "step": 9 }, { "epoch": 0.0022647491790284227, "grad_norm": 0.40149691700935364, "learning_rate": 0.0002, "loss": 0.3227, "step": 10 } ], "logging_steps": 1, "max_steps": 10, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 3, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 3822182832537600.0, "train_batch_size": 2, "trial_name": null, "trial_params": null }