{ "best_metric": null, "best_model_checkpoint": null, "epoch": 38.46153846153846, "eval_steps": 500, "global_step": 2500, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 7.6923076923076925, "grad_norm": 103.745361328125, "learning_rate": 9.950700000000001e-06, "loss": 3.7724, "step": 500 }, { "epoch": 7.6923076923076925, "eval_loss": 4.843122959136963, "eval_runtime": 12.2345, "eval_samples_per_second": 10.707, "eval_steps_per_second": 1.39, "step": 500 }, { "epoch": 15.384615384615385, "grad_norm": 134.67405700683594, "learning_rate": 9.9008e-06, "loss": 2.823, "step": 1000 }, { "epoch": 15.384615384615385, "eval_loss": 4.731071472167969, "eval_runtime": 11.7452, "eval_samples_per_second": 11.153, "eval_steps_per_second": 1.447, "step": 1000 }, { "epoch": 23.076923076923077, "grad_norm": 67.87723541259766, "learning_rate": 9.8508e-06, "loss": 2.6546, "step": 1500 }, { "epoch": 23.076923076923077, "eval_loss": 4.85584020614624, "eval_runtime": 12.0536, "eval_samples_per_second": 10.868, "eval_steps_per_second": 1.41, "step": 1500 }, { "epoch": 30.76923076923077, "grad_norm": 60.380435943603516, "learning_rate": 9.800800000000001e-06, "loss": 2.4997, "step": 2000 }, { "epoch": 30.76923076923077, "eval_loss": 4.691804885864258, "eval_runtime": 12.3485, "eval_samples_per_second": 10.609, "eval_steps_per_second": 1.377, "step": 2000 }, { "epoch": 38.46153846153846, "grad_norm": 70.92816925048828, "learning_rate": 9.7508e-06, "loss": 2.3852, "step": 2500 }, { "epoch": 38.46153846153846, "eval_loss": 4.495730400085449, "eval_runtime": 11.9295, "eval_samples_per_second": 10.981, "eval_steps_per_second": 1.425, "step": 2500 } ], "logging_steps": 500, "max_steps": 100000, "num_input_tokens_seen": 0, "num_train_epochs": 1539, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 9.5560791552e+18, "train_batch_size": 8, "trial_name": null, "trial_params": null }