{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.00899685110211426, "eval_steps": 3, "global_step": 10, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.000899685110211426, "grad_norm": 0.950228214263916, "learning_rate": 2e-05, "loss": 4.87, "step": 1 }, { "epoch": 0.000899685110211426, "eval_loss": 1.221696138381958, "eval_runtime": 45.5017, "eval_samples_per_second": 10.307, "eval_steps_per_second": 5.165, "step": 1 }, { "epoch": 0.001799370220422852, "grad_norm": 1.0775448083877563, "learning_rate": 4e-05, "loss": 4.4487, "step": 2 }, { "epoch": 0.002699055330634278, "grad_norm": 0.8101547360420227, "learning_rate": 6e-05, "loss": 4.6589, "step": 3 }, { "epoch": 0.002699055330634278, "eval_loss": 1.220987319946289, "eval_runtime": 45.5214, "eval_samples_per_second": 10.303, "eval_steps_per_second": 5.162, "step": 3 }, { "epoch": 0.003598740440845704, "grad_norm": 0.8206847310066223, "learning_rate": 8e-05, "loss": 4.6813, "step": 4 }, { "epoch": 0.00449842555105713, "grad_norm": 1.1539366245269775, "learning_rate": 0.0001, "loss": 5.1025, "step": 5 }, { "epoch": 0.005398110661268556, "grad_norm": 1.1764967441558838, "learning_rate": 0.00012, "loss": 4.0469, "step": 6 }, { "epoch": 0.005398110661268556, "eval_loss": 1.2108772993087769, "eval_runtime": 45.631, "eval_samples_per_second": 10.278, "eval_steps_per_second": 5.15, "step": 6 }, { "epoch": 0.006297795771479982, "grad_norm": 1.6356772184371948, "learning_rate": 0.00014, "loss": 4.8693, "step": 7 }, { "epoch": 0.007197480881691408, "grad_norm": 1.2193690538406372, "learning_rate": 0.00016, "loss": 4.3444, "step": 8 }, { "epoch": 0.008097165991902834, "grad_norm": 3.0683789253234863, "learning_rate": 0.00018, "loss": 7.3665, "step": 9 }, { "epoch": 0.008097165991902834, "eval_loss": 1.172175645828247, "eval_runtime": 45.5489, "eval_samples_per_second": 10.297, "eval_steps_per_second": 5.159, "step": 9 }, { "epoch": 0.00899685110211426, "grad_norm": 1.8598663806915283, "learning_rate": 0.0002, "loss": 4.2505, "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": 5605656153292800.0, "train_batch_size": 2, "trial_name": null, "trial_params": null }