{ "best_metric": 10.352457046508789, "best_model_checkpoint": "miner_id_24/checkpoint-50", "epoch": 0.857449088960343, "eval_steps": 25, "global_step": 50, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.01714898177920686, "grad_norm": 0.2630438804626465, "learning_rate": 5e-05, "loss": 10.3751, "step": 1 }, { "epoch": 0.01714898177920686, "eval_loss": 10.376319885253906, "eval_runtime": 0.1255, "eval_samples_per_second": 398.25, "eval_steps_per_second": 103.545, "step": 1 }, { "epoch": 0.03429796355841372, "grad_norm": 0.25066354870796204, "learning_rate": 0.0001, "loss": 10.3769, "step": 2 }, { "epoch": 0.05144694533762058, "grad_norm": 0.22949962317943573, "learning_rate": 9.990365154573717e-05, "loss": 10.3757, "step": 3 }, { "epoch": 0.06859592711682744, "grad_norm": 0.22806420922279358, "learning_rate": 9.961501876182148e-05, "loss": 10.3754, "step": 4 }, { "epoch": 0.0857449088960343, "grad_norm": 0.24396614730358124, "learning_rate": 9.913533761814537e-05, "loss": 10.3758, "step": 5 }, { "epoch": 0.10289389067524116, "grad_norm": 0.2243625670671463, "learning_rate": 9.846666218300807e-05, "loss": 10.3737, "step": 6 }, { "epoch": 0.12004287245444802, "grad_norm": 0.22151897847652435, "learning_rate": 9.761185582727977e-05, "loss": 10.3736, "step": 7 }, { "epoch": 0.13719185423365488, "grad_norm": 0.20165084302425385, "learning_rate": 9.657457896300791e-05, "loss": 10.3721, "step": 8 }, { "epoch": 0.15434083601286175, "grad_norm": 0.13735997676849365, "learning_rate": 9.535927336897098e-05, "loss": 10.3749, "step": 9 }, { "epoch": 0.1714898177920686, "grad_norm": 0.14907795190811157, "learning_rate": 9.397114317029975e-05, "loss": 10.3736, "step": 10 }, { "epoch": 0.18863879957127547, "grad_norm": 0.1417752057313919, "learning_rate": 9.241613255361455e-05, "loss": 10.3708, "step": 11 }, { "epoch": 0.2057877813504823, "grad_norm": 0.16261570155620575, "learning_rate": 9.070090031310558e-05, "loss": 10.3705, "step": 12 }, { "epoch": 0.22293676312968919, "grad_norm": 0.2875140905380249, "learning_rate": 8.883279133655399e-05, "loss": 10.3654, "step": 13 }, { "epoch": 0.24008574490889603, "grad_norm": 0.2875608205795288, "learning_rate": 8.681980515339464e-05, "loss": 10.365, "step": 14 }, { "epoch": 0.2572347266881029, "grad_norm": 0.2791755795478821, "learning_rate": 8.467056167950311e-05, "loss": 10.3684, "step": 15 }, { "epoch": 0.27438370846730975, "grad_norm": 0.2789701223373413, "learning_rate": 8.239426430539243e-05, "loss": 10.3667, "step": 16 }, { "epoch": 0.2915326902465166, "grad_norm": 0.30052244663238525, "learning_rate": 8.000066048588211e-05, "loss": 10.3647, "step": 17 }, { "epoch": 0.3086816720257235, "grad_norm": 0.2686644494533539, "learning_rate": 7.75e-05, "loss": 10.3647, "step": 18 }, { "epoch": 0.3258306538049303, "grad_norm": 0.28808850049972534, "learning_rate": 7.490299105985507e-05, "loss": 10.3621, "step": 19 }, { "epoch": 0.3429796355841372, "grad_norm": 0.2524069845676422, "learning_rate": 7.222075445642904e-05, "loss": 10.3644, "step": 20 }, { "epoch": 0.36012861736334406, "grad_norm": 0.21884575486183167, "learning_rate": 6.946477593864228e-05, "loss": 10.3652, "step": 21 }, { "epoch": 0.37727759914255093, "grad_norm": 0.18064546585083008, "learning_rate": 6.664685702961344e-05, "loss": 10.3667, "step": 22 }, { "epoch": 0.39442658092175775, "grad_norm": 0.1982426643371582, "learning_rate": 6.377906449072578e-05, "loss": 10.3634, "step": 23 }, { "epoch": 0.4115755627009646, "grad_norm": 0.20135067403316498, "learning_rate": 6.087367864990233e-05, "loss": 10.3649, "step": 24 }, { "epoch": 0.4287245444801715, "grad_norm": 0.2692125141620636, "learning_rate": 5.794314081535644e-05, "loss": 10.3577, "step": 25 }, { "epoch": 0.4287245444801715, "eval_loss": 10.3605375289917, "eval_runtime": 0.129, "eval_samples_per_second": 387.662, "eval_steps_per_second": 100.792, "step": 25 }, { "epoch": 0.44587352625937837, "grad_norm": 0.34039604663848877, "learning_rate": 5.500000000000001e-05, "loss": 10.3547, "step": 26 }, { "epoch": 0.4630225080385852, "grad_norm": 0.3508872091770172, "learning_rate": 5.205685918464356e-05, "loss": 10.3533, "step": 27 }, { "epoch": 0.48017148981779206, "grad_norm": 0.3430570065975189, "learning_rate": 4.912632135009769e-05, "loss": 10.3551, "step": 28 }, { "epoch": 0.49732047159699894, "grad_norm": 0.3169425129890442, "learning_rate": 4.6220935509274235e-05, "loss": 10.3567, "step": 29 }, { "epoch": 0.5144694533762058, "grad_norm": 0.34082862734794617, "learning_rate": 4.3353142970386564e-05, "loss": 10.3553, "step": 30 }, { "epoch": 0.5316184351554126, "grad_norm": 0.32129889726638794, "learning_rate": 4.053522406135775e-05, "loss": 10.3557, "step": 31 }, { "epoch": 0.5487674169346195, "grad_norm": 0.3115546703338623, "learning_rate": 3.777924554357096e-05, "loss": 10.3563, "step": 32 }, { "epoch": 0.5659163987138264, "grad_norm": 0.2841951847076416, "learning_rate": 3.509700894014496e-05, "loss": 10.3599, "step": 33 }, { "epoch": 0.5830653804930332, "grad_norm": 0.24529126286506653, "learning_rate": 3.250000000000001e-05, "loss": 10.3589, "step": 34 }, { "epoch": 0.6002143622722401, "grad_norm": 0.22551962733268738, "learning_rate": 2.9999339514117912e-05, "loss": 10.3594, "step": 35 }, { "epoch": 0.617363344051447, "grad_norm": 0.19762487709522247, "learning_rate": 2.760573569460757e-05, "loss": 10.361, "step": 36 }, { "epoch": 0.6345123258306538, "grad_norm": 0.21796438097953796, "learning_rate": 2.53294383204969e-05, "loss": 10.3592, "step": 37 }, { "epoch": 0.6516613076098606, "grad_norm": 0.38318026065826416, "learning_rate": 2.3180194846605367e-05, "loss": 10.3482, "step": 38 }, { "epoch": 0.6688102893890675, "grad_norm": 0.3688681721687317, "learning_rate": 2.1167208663446025e-05, "loss": 10.3481, "step": 39 }, { "epoch": 0.6859592711682744, "grad_norm": 0.3773597478866577, "learning_rate": 1.9299099686894423e-05, "loss": 10.3506, "step": 40 }, { "epoch": 0.7031082529474812, "grad_norm": 0.3649727404117584, "learning_rate": 1.758386744638546e-05, "loss": 10.3506, "step": 41 }, { "epoch": 0.7202572347266881, "grad_norm": 0.38561907410621643, "learning_rate": 1.602885682970026e-05, "loss": 10.3485, "step": 42 }, { "epoch": 0.737406216505895, "grad_norm": 0.3504658043384552, "learning_rate": 1.464072663102903e-05, "loss": 10.3508, "step": 43 }, { "epoch": 0.7545551982851019, "grad_norm": 0.3569239377975464, "learning_rate": 1.3425421036992098e-05, "loss": 10.3515, "step": 44 }, { "epoch": 0.7717041800643086, "grad_norm": 0.30078503489494324, "learning_rate": 1.2388144172720251e-05, "loss": 10.3536, "step": 45 }, { "epoch": 0.7888531618435155, "grad_norm": 0.2712373733520508, "learning_rate": 1.1533337816991932e-05, "loss": 10.3563, "step": 46 }, { "epoch": 0.8060021436227224, "grad_norm": 0.21626907587051392, "learning_rate": 1.0864662381854632e-05, "loss": 10.3594, "step": 47 }, { "epoch": 0.8231511254019293, "grad_norm": 0.2447448968887329, "learning_rate": 1.0384981238178534e-05, "loss": 10.3575, "step": 48 }, { "epoch": 0.8403001071811361, "grad_norm": 0.2010498046875, "learning_rate": 1.0096348454262845e-05, "loss": 10.3599, "step": 49 }, { "epoch": 0.857449088960343, "grad_norm": 0.24134501814842224, "learning_rate": 1e-05, "loss": 10.3564, "step": 50 }, { "epoch": 0.857449088960343, "eval_loss": 10.352457046508789, "eval_runtime": 0.1235, "eval_samples_per_second": 404.722, "eval_steps_per_second": 105.228, "step": 50 } ], "logging_steps": 1, "max_steps": 50, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 25, "stateful_callbacks": { "EarlyStoppingCallback": { "args": { "early_stopping_patience": 1, "early_stopping_threshold": 0.0 }, "attributes": { "early_stopping_patience_counter": 0 } }, "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 51188021329920.0, "train_batch_size": 2, "trial_name": null, "trial_params": null }