{ "best_metric": null, "best_model_checkpoint": null, "epoch": 1.6889238020424195, "eval_steps": 50, "global_step": 2150, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.03927729772191673, "grad_norm": 0.3003779351711273, "learning_rate": 6.544502617801048e-05, "loss": 4.2186, "step": 50 }, { "epoch": 0.03927729772191673, "eval_loss": 3.861752510070801, "eval_runtime": 29.4492, "eval_samples_per_second": 3.396, "eval_steps_per_second": 0.441, "step": 50 }, { "epoch": 0.07855459544383346, "grad_norm": 0.19962768256664276, "learning_rate": 0.00013089005235602096, "loss": 3.1772, "step": 100 }, { "epoch": 0.07855459544383346, "eval_loss": 2.59340238571167, "eval_runtime": 29.4336, "eval_samples_per_second": 3.397, "eval_steps_per_second": 0.442, "step": 100 }, { "epoch": 0.1178318931657502, "grad_norm": 0.17451965808868408, "learning_rate": 0.00019633507853403142, "loss": 2.401, "step": 150 }, { "epoch": 0.1178318931657502, "eval_loss": 2.2559566497802734, "eval_runtime": 29.4433, "eval_samples_per_second": 3.396, "eval_steps_per_second": 0.442, "step": 150 }, { "epoch": 0.15710919088766692, "grad_norm": 0.18111871182918549, "learning_rate": 0.0002617801047120419, "loss": 2.1397, "step": 200 }, { "epoch": 0.15710919088766692, "eval_loss": 2.1369433403015137, "eval_runtime": 29.6429, "eval_samples_per_second": 3.373, "eval_steps_per_second": 0.439, "step": 200 }, { "epoch": 0.19638648860958366, "grad_norm": 0.18973998725414276, "learning_rate": 0.0003272251308900523, "loss": 2.0834, "step": 250 }, { "epoch": 0.19638648860958366, "eval_loss": 2.08046555519104, "eval_runtime": 29.5575, "eval_samples_per_second": 3.383, "eval_steps_per_second": 0.44, "step": 250 }, { "epoch": 0.2356637863315004, "grad_norm": 0.1734466403722763, "learning_rate": 0.00039267015706806284, "loss": 2.055, "step": 300 }, { "epoch": 0.2356637863315004, "eval_loss": 2.056304693222046, "eval_runtime": 29.4932, "eval_samples_per_second": 3.391, "eval_steps_per_second": 0.441, "step": 300 }, { "epoch": 0.2749410840534171, "grad_norm": 0.21972551941871643, "learning_rate": 0.0004581151832460733, "loss": 2.043, "step": 350 }, { "epoch": 0.2749410840534171, "eval_loss": 2.0286335945129395, "eval_runtime": 29.3134, "eval_samples_per_second": 3.411, "eval_steps_per_second": 0.443, "step": 350 }, { "epoch": 0.31421838177533384, "grad_norm": 0.20237264037132263, "learning_rate": 0.000497381437299971, "loss": 2.0135, "step": 400 }, { "epoch": 0.31421838177533384, "eval_loss": 2.0177054405212402, "eval_runtime": 29.516, "eval_samples_per_second": 3.388, "eval_steps_per_second": 0.44, "step": 400 }, { "epoch": 0.3534956794972506, "grad_norm": 0.18964040279388428, "learning_rate": 0.0004901076520221123, "loss": 1.9971, "step": 450 }, { "epoch": 0.3534956794972506, "eval_loss": 2.0019736289978027, "eval_runtime": 29.578, "eval_samples_per_second": 3.381, "eval_steps_per_second": 0.44, "step": 450 }, { "epoch": 0.3927729772191673, "grad_norm": 0.20875422656536102, "learning_rate": 0.00048283386674425374, "loss": 1.9766, "step": 500 }, { "epoch": 0.3927729772191673, "eval_loss": 1.991439700126648, "eval_runtime": 29.3577, "eval_samples_per_second": 3.406, "eval_steps_per_second": 0.443, "step": 500 }, { "epoch": 0.43205027494108406, "grad_norm": 0.1806900054216385, "learning_rate": 0.00047556008146639513, "loss": 1.9677, "step": 550 }, { "epoch": 0.43205027494108406, "eval_loss": 1.9789185523986816, "eval_runtime": 29.5883, "eval_samples_per_second": 3.38, "eval_steps_per_second": 0.439, "step": 550 }, { "epoch": 0.4713275726630008, "grad_norm": 0.17965032160282135, "learning_rate": 0.00046828629618853653, "loss": 1.9562, "step": 600 }, { "epoch": 0.4713275726630008, "eval_loss": 1.9679882526397705, "eval_runtime": 29.7136, "eval_samples_per_second": 3.365, "eval_steps_per_second": 0.438, "step": 600 }, { "epoch": 0.5106048703849175, "grad_norm": 0.16700418293476105, "learning_rate": 0.0004610125109106779, "loss": 1.9594, "step": 650 }, { "epoch": 0.5106048703849175, "eval_loss": 1.9631274938583374, "eval_runtime": 29.8472, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.436, "step": 650 }, { "epoch": 0.5498821681068342, "grad_norm": 0.1479049026966095, "learning_rate": 0.0004537387256328193, "loss": 1.9423, "step": 700 }, { "epoch": 0.5498821681068342, "eval_loss": 1.9546252489089966, "eval_runtime": 29.5017, "eval_samples_per_second": 3.39, "eval_steps_per_second": 0.441, "step": 700 }, { "epoch": 0.589159465828751, "grad_norm": 0.14641450345516205, "learning_rate": 0.0004464649403549607, "loss": 1.9587, "step": 750 }, { "epoch": 0.589159465828751, "eval_loss": 1.9469780921936035, "eval_runtime": 29.8753, "eval_samples_per_second": 3.347, "eval_steps_per_second": 0.435, "step": 750 }, { "epoch": 0.6284367635506677, "grad_norm": 0.19608749449253082, "learning_rate": 0.0004391911550771021, "loss": 1.9408, "step": 800 }, { "epoch": 0.6284367635506677, "eval_loss": 1.9397352933883667, "eval_runtime": 29.4107, "eval_samples_per_second": 3.4, "eval_steps_per_second": 0.442, "step": 800 }, { "epoch": 0.6677140612725845, "grad_norm": 0.18671496212482452, "learning_rate": 0.0004319173697992435, "loss": 1.9816, "step": 850 }, { "epoch": 0.6677140612725845, "eval_loss": 1.9425079822540283, "eval_runtime": 29.4898, "eval_samples_per_second": 3.391, "eval_steps_per_second": 0.441, "step": 850 }, { "epoch": 0.7069913589945012, "grad_norm": 0.16066303849220276, "learning_rate": 0.00042464358452138495, "loss": 1.9298, "step": 900 }, { "epoch": 0.7069913589945012, "eval_loss": 1.9176784753799438, "eval_runtime": 29.544, "eval_samples_per_second": 3.385, "eval_steps_per_second": 0.44, "step": 900 }, { "epoch": 0.746268656716418, "grad_norm": 0.17130565643310547, "learning_rate": 0.00041736979924352635, "loss": 1.9021, "step": 950 }, { "epoch": 0.746268656716418, "eval_loss": 1.9149932861328125, "eval_runtime": 29.4076, "eval_samples_per_second": 3.4, "eval_steps_per_second": 0.442, "step": 950 }, { "epoch": 0.7855459544383346, "grad_norm": 0.15130634605884552, "learning_rate": 0.00041009601396566775, "loss": 1.9104, "step": 1000 }, { "epoch": 0.7855459544383346, "eval_loss": 1.9071648120880127, "eval_runtime": 29.4527, "eval_samples_per_second": 3.395, "eval_steps_per_second": 0.441, "step": 1000 }, { "epoch": 0.8248232521602514, "grad_norm": 0.1651030331850052, "learning_rate": 0.00040282222868780914, "loss": 1.9325, "step": 1050 }, { "epoch": 0.8248232521602514, "eval_loss": 1.8993133306503296, "eval_runtime": 29.5512, "eval_samples_per_second": 3.384, "eval_steps_per_second": 0.44, "step": 1050 }, { "epoch": 0.8641005498821681, "grad_norm": 0.16182482242584229, "learning_rate": 0.00039554844340995054, "loss": 1.9183, "step": 1100 }, { "epoch": 0.8641005498821681, "eval_loss": 1.9054021835327148, "eval_runtime": 29.3876, "eval_samples_per_second": 3.403, "eval_steps_per_second": 0.442, "step": 1100 }, { "epoch": 0.9033778476040848, "grad_norm": 0.16718855500221252, "learning_rate": 0.00038827465813209193, "loss": 1.9557, "step": 1150 }, { "epoch": 0.9033778476040848, "eval_loss": 1.894838571548462, "eval_runtime": 29.428, "eval_samples_per_second": 3.398, "eval_steps_per_second": 0.442, "step": 1150 }, { "epoch": 0.9426551453260016, "grad_norm": 0.14458701014518738, "learning_rate": 0.0003810008728542333, "loss": 1.9261, "step": 1200 }, { "epoch": 0.9426551453260016, "eval_loss": 1.8822929859161377, "eval_runtime": 29.4891, "eval_samples_per_second": 3.391, "eval_steps_per_second": 0.441, "step": 1200 }, { "epoch": 0.9819324430479183, "grad_norm": 0.1621202528476715, "learning_rate": 0.0003737270875763747, "loss": 1.9337, "step": 1250 }, { "epoch": 0.9819324430479183, "eval_loss": 1.8784922361373901, "eval_runtime": 29.4313, "eval_samples_per_second": 3.398, "eval_steps_per_second": 0.442, "step": 1250 }, { "epoch": 1.021209740769835, "grad_norm": 0.17562049627304077, "learning_rate": 0.00036645330229851617, "loss": 1.9034, "step": 1300 }, { "epoch": 1.021209740769835, "eval_loss": 1.8769853115081787, "eval_runtime": 29.4816, "eval_samples_per_second": 3.392, "eval_steps_per_second": 0.441, "step": 1300 }, { "epoch": 1.0604870384917517, "grad_norm": 0.18526038527488708, "learning_rate": 0.00035917951702065757, "loss": 1.8603, "step": 1350 }, { "epoch": 1.0604870384917517, "eval_loss": 1.8668268918991089, "eval_runtime": 29.5392, "eval_samples_per_second": 3.385, "eval_steps_per_second": 0.44, "step": 1350 }, { "epoch": 1.0997643362136684, "grad_norm": 0.1711549311876297, "learning_rate": 0.00035190573174279896, "loss": 1.8477, "step": 1400 }, { "epoch": 1.0997643362136684, "eval_loss": 1.8661786317825317, "eval_runtime": 29.4511, "eval_samples_per_second": 3.395, "eval_steps_per_second": 0.441, "step": 1400 }, { "epoch": 1.139041633935585, "grad_norm": 0.18905235826969147, "learning_rate": 0.00034463194646494036, "loss": 1.8658, "step": 1450 }, { "epoch": 1.139041633935585, "eval_loss": 1.8573988676071167, "eval_runtime": 29.4384, "eval_samples_per_second": 3.397, "eval_steps_per_second": 0.442, "step": 1450 }, { "epoch": 1.178318931657502, "grad_norm": 0.16769160330295563, "learning_rate": 0.0003373581611870818, "loss": 1.8923, "step": 1500 }, { "epoch": 1.178318931657502, "eval_loss": 1.857424259185791, "eval_runtime": 29.4845, "eval_samples_per_second": 3.392, "eval_steps_per_second": 0.441, "step": 1500 }, { "epoch": 1.2175962293794187, "grad_norm": 0.2095043659210205, "learning_rate": 0.0003300843759092232, "loss": 1.8777, "step": 1550 }, { "epoch": 1.2175962293794187, "eval_loss": 1.8602547645568848, "eval_runtime": 29.5084, "eval_samples_per_second": 3.389, "eval_steps_per_second": 0.441, "step": 1550 }, { "epoch": 1.2568735271013354, "grad_norm": 0.18356919288635254, "learning_rate": 0.0003228105906313646, "loss": 1.8645, "step": 1600 }, { "epoch": 1.2568735271013354, "eval_loss": 1.8516952991485596, "eval_runtime": 29.539, "eval_samples_per_second": 3.385, "eval_steps_per_second": 0.44, "step": 1600 }, { "epoch": 1.2961508248232523, "grad_norm": 0.23483967781066895, "learning_rate": 0.000315536805353506, "loss": 1.8204, "step": 1650 }, { "epoch": 1.2961508248232523, "eval_loss": 1.8447459936141968, "eval_runtime": 29.4621, "eval_samples_per_second": 3.394, "eval_steps_per_second": 0.441, "step": 1650 }, { "epoch": 1.335428122545169, "grad_norm": 0.21005484461784363, "learning_rate": 0.0003082630200756474, "loss": 1.8661, "step": 1700 }, { "epoch": 1.335428122545169, "eval_loss": 1.8399604558944702, "eval_runtime": 29.4709, "eval_samples_per_second": 3.393, "eval_steps_per_second": 0.441, "step": 1700 }, { "epoch": 1.3747054202670856, "grad_norm": 0.19819919764995575, "learning_rate": 0.0003009892347977888, "loss": 1.8595, "step": 1750 }, { "epoch": 1.3747054202670856, "eval_loss": 1.8384497165679932, "eval_runtime": 29.4803, "eval_samples_per_second": 3.392, "eval_steps_per_second": 0.441, "step": 1750 }, { "epoch": 1.4139827179890023, "grad_norm": 0.21313180029392242, "learning_rate": 0.0002937154495199302, "loss": 1.857, "step": 1800 }, { "epoch": 1.4139827179890023, "eval_loss": 1.831393837928772, "eval_runtime": 29.4965, "eval_samples_per_second": 3.39, "eval_steps_per_second": 0.441, "step": 1800 }, { "epoch": 1.453260015710919, "grad_norm": 0.2181958109140396, "learning_rate": 0.0002864416642420716, "loss": 1.8431, "step": 1850 }, { "epoch": 1.453260015710919, "eval_loss": 1.827854037284851, "eval_runtime": 29.4535, "eval_samples_per_second": 3.395, "eval_steps_per_second": 0.441, "step": 1850 }, { "epoch": 1.4925373134328357, "grad_norm": 0.2101767212152481, "learning_rate": 0.00027916787896421297, "loss": 1.8249, "step": 1900 }, { "epoch": 1.4925373134328357, "eval_loss": 1.8285325765609741, "eval_runtime": 29.5818, "eval_samples_per_second": 3.38, "eval_steps_per_second": 0.439, "step": 1900 }, { "epoch": 1.5318146111547526, "grad_norm": 0.23812232911586761, "learning_rate": 0.00027189409368635437, "loss": 1.8372, "step": 1950 }, { "epoch": 1.5318146111547526, "eval_loss": 1.824296474456787, "eval_runtime": 29.4834, "eval_samples_per_second": 3.392, "eval_steps_per_second": 0.441, "step": 1950 }, { "epoch": 1.5710919088766693, "grad_norm": 0.1917562186717987, "learning_rate": 0.00026462030840849576, "loss": 1.8589, "step": 2000 }, { "epoch": 1.5710919088766693, "eval_loss": 1.8210265636444092, "eval_runtime": 29.5979, "eval_samples_per_second": 3.379, "eval_steps_per_second": 0.439, "step": 2000 }, { "epoch": 1.6103692065985862, "grad_norm": 0.222442165017128, "learning_rate": 0.00025734652313063716, "loss": 1.829, "step": 2050 }, { "epoch": 1.6103692065985862, "eval_loss": 1.8052531480789185, "eval_runtime": 29.6699, "eval_samples_per_second": 3.37, "eval_steps_per_second": 0.438, "step": 2050 }, { "epoch": 1.6496465043205029, "grad_norm": 0.19593429565429688, "learning_rate": 0.00025007273785277855, "loss": 1.8154, "step": 2100 }, { "epoch": 1.6496465043205029, "eval_loss": 1.800155758857727, "eval_runtime": 29.4349, "eval_samples_per_second": 3.397, "eval_steps_per_second": 0.442, "step": 2100 } ], "logging_steps": 50, "max_steps": 3819, "num_input_tokens_seen": 0, "num_train_epochs": 3, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 2.643342548957184e+16, "train_batch_size": 8, "trial_name": null, "trial_params": null }