{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.046189376443418015, "eval_steps": 500, "global_step": 20, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0, "learning_rate": 2e-05, "loss": 1.6463, "step": 1 }, { "epoch": 0.0, "learning_rate": 4e-05, "loss": 4.8805, "step": 2 }, { "epoch": 0.01, "learning_rate": 6e-05, "loss": 4.4954, "step": 3 }, { "epoch": 0.01, "learning_rate": 8e-05, "loss": 4.4632, "step": 4 }, { "epoch": 0.01, "learning_rate": 0.0001, "loss": 4.8846, "step": 5 }, { "epoch": 0.01, "learning_rate": 0.00012, "loss": 5.0885, "step": 6 }, { "epoch": 0.02, "learning_rate": 0.00014, "loss": 3.0114, "step": 7 }, { "epoch": 0.02, "learning_rate": 0.00016, "loss": 4.455, "step": 8 }, { "epoch": 0.02, "learning_rate": 0.00018, "loss": 2.209, "step": 9 }, { "epoch": 0.02, "learning_rate": 0.0002, "loss": 4.4377, "step": 10 }, { "epoch": 0.03, "learning_rate": 0.00019510565162951537, "loss": 3.7615, "step": 11 }, { "epoch": 0.03, "learning_rate": 0.00018090169943749476, "loss": 3.5525, "step": 12 }, { "epoch": 0.03, "learning_rate": 0.00015877852522924732, "loss": 3.7633, "step": 13 }, { "epoch": 0.03, "learning_rate": 0.00013090169943749476, "loss": 3.5811, "step": 14 }, { "epoch": 0.03, "learning_rate": 0.0001, "loss": 3.3589, "step": 15 }, { "epoch": 0.04, "learning_rate": 6.909830056250527e-05, "loss": 3.7917, "step": 16 }, { "epoch": 0.04, "learning_rate": 4.12214747707527e-05, "loss": 3.4175, "step": 17 }, { "epoch": 0.04, "learning_rate": 1.9098300562505266e-05, "loss": 3.7471, "step": 18 }, { "epoch": 0.04, "learning_rate": 4.8943483704846475e-06, "loss": 3.6897, "step": 19 }, { "epoch": 0.05, "learning_rate": 0.0, "loss": 4.0306, "step": 20 }, { "epoch": 0.05, "eval_loss": 3.5158474445343018, "eval_runtime": 5.5975, "eval_samples_per_second": 17.865, "eval_steps_per_second": 17.865, "step": 20 } ], "logging_steps": 1, "max_steps": 20, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 500, "total_flos": 146498975170560.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }