{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.01794168950909544, "eval_steps": 9, "global_step": 54, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.00033225350942769335, "eval_loss": 2.1395199298858643, "eval_runtime": 220.9563, "eval_samples_per_second": 22.941, "eval_steps_per_second": 2.869, "step": 1 }, { "epoch": 0.00099676052828308, "grad_norm": 5.9584269523620605, "learning_rate": 1.5e-05, "loss": 9.4483, "step": 3 }, { "epoch": 0.00199352105656616, "grad_norm": 4.213073253631592, "learning_rate": 3e-05, "loss": 7.8813, "step": 6 }, { "epoch": 0.00299028158484924, "grad_norm": 4.823021411895752, "learning_rate": 4.5e-05, "loss": 7.9607, "step": 9 }, { "epoch": 0.00299028158484924, "eval_loss": 2.070302963256836, "eval_runtime": 221.304, "eval_samples_per_second": 22.905, "eval_steps_per_second": 2.865, "step": 9 }, { "epoch": 0.00398704211313232, "grad_norm": 7.021877765655518, "learning_rate": 4.993910125649561e-05, "loss": 9.0723, "step": 12 }, { "epoch": 0.0049838026414154, "grad_norm": 10.367716789245605, "learning_rate": 4.962019382530521e-05, "loss": 8.6656, "step": 15 }, { "epoch": 0.00598056316969848, "grad_norm": 3.396527051925659, "learning_rate": 4.9031542398457974e-05, "loss": 6.5636, "step": 18 }, { "epoch": 0.00598056316969848, "eval_loss": 1.8593316078186035, "eval_runtime": 221.3173, "eval_samples_per_second": 22.904, "eval_steps_per_second": 2.865, "step": 18 }, { "epoch": 0.00697732369798156, "grad_norm": 5.275881290435791, "learning_rate": 4.817959636416969e-05, "loss": 7.3881, "step": 21 }, { "epoch": 0.00797408422626464, "grad_norm": 4.454370975494385, "learning_rate": 4.707368982147318e-05, "loss": 7.0281, "step": 24 }, { "epoch": 0.00897084475454772, "grad_norm": 4.078393936157227, "learning_rate": 4.572593931387604e-05, "loss": 6.9298, "step": 27 }, { "epoch": 0.00897084475454772, "eval_loss": 1.7337881326675415, "eval_runtime": 221.3769, "eval_samples_per_second": 22.898, "eval_steps_per_second": 2.864, "step": 27 }, { "epoch": 0.0099676052828308, "grad_norm": 4.427830696105957, "learning_rate": 4.415111107797445e-05, "loss": 6.7986, "step": 30 }, { "epoch": 0.01096436581111388, "grad_norm": 4.341389179229736, "learning_rate": 4.2366459261474933e-05, "loss": 6.6719, "step": 33 }, { "epoch": 0.01196112633939696, "grad_norm": 6.339109420776367, "learning_rate": 4.039153688314145e-05, "loss": 8.2781, "step": 36 }, { "epoch": 0.01196112633939696, "eval_loss": 1.6621569395065308, "eval_runtime": 221.3208, "eval_samples_per_second": 22.903, "eval_steps_per_second": 2.865, "step": 36 }, { "epoch": 0.01295788686768004, "grad_norm": 4.891085147857666, "learning_rate": 3.824798160583012e-05, "loss": 6.0951, "step": 39 }, { "epoch": 0.01395464739596312, "grad_norm": 4.013057231903076, "learning_rate": 3.5959278669726935e-05, "loss": 6.5056, "step": 42 }, { "epoch": 0.0149514079242462, "grad_norm": 3.513582468032837, "learning_rate": 3.355050358314172e-05, "loss": 6.14, "step": 45 }, { "epoch": 0.0149514079242462, "eval_loss": 1.614381194114685, "eval_runtime": 221.2948, "eval_samples_per_second": 22.906, "eval_steps_per_second": 2.865, "step": 45 }, { "epoch": 0.01594816845252928, "grad_norm": 4.265641212463379, "learning_rate": 3.104804738999169e-05, "loss": 4.8365, "step": 48 }, { "epoch": 0.01694492898081236, "grad_norm": 4.8743062019348145, "learning_rate": 2.8479327524001636e-05, "loss": 7.3449, "step": 51 }, { "epoch": 0.01794168950909544, "grad_norm": 4.481264591217041, "learning_rate": 2.587248741756253e-05, "loss": 6.5566, "step": 54 }, { "epoch": 0.01794168950909544, "eval_loss": 1.5844690799713135, "eval_runtime": 221.4081, "eval_samples_per_second": 22.894, "eval_steps_per_second": 2.863, "step": 54 } ], "logging_steps": 3, "max_steps": 100, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 9, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 4505059178053632.0, "train_batch_size": 8, "trial_name": null, "trial_params": null }