| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.32923775694404, | |
| "eval_steps": 10000, | |
| "global_step": 40000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0291154719618005, | |
| "grad_norm": 5.6875, | |
| "learning_rate": 4.9514742133969994e-05, | |
| "loss": 2.4005, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.058230943923601, | |
| "grad_norm": 5.125, | |
| "learning_rate": 4.9029484267939986e-05, | |
| "loss": 1.8762, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.0873464158854015, | |
| "grad_norm": 5.21875, | |
| "learning_rate": 4.854422640190998e-05, | |
| "loss": 1.7508, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.116461887847202, | |
| "grad_norm": 5.125, | |
| "learning_rate": 4.805896853587997e-05, | |
| "loss": 1.6692, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.1455773598090025, | |
| "grad_norm": 5.03125, | |
| "learning_rate": 4.757371066984996e-05, | |
| "loss": 1.6309, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.174692831770803, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 4.708845280381995e-05, | |
| "loss": 1.5878, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.2038083037326035, | |
| "grad_norm": 4.90625, | |
| "learning_rate": 4.6603194937789944e-05, | |
| "loss": 1.5593, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.232923775694404, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 4.6117937071759936e-05, | |
| "loss": 1.5287, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.2620392476562045, | |
| "grad_norm": 5.53125, | |
| "learning_rate": 4.563267920572993e-05, | |
| "loss": 1.5189, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.291154719618005, | |
| "grad_norm": 4.53125, | |
| "learning_rate": 4.514742133969992e-05, | |
| "loss": 1.5018, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.3202701915798055, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 4.466216347366991e-05, | |
| "loss": 1.4784, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.349385663541606, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 4.41769056076399e-05, | |
| "loss": 1.4703, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.3785011355034065, | |
| "grad_norm": 4.5, | |
| "learning_rate": 4.3691647741609894e-05, | |
| "loss": 1.449, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.407616607465207, | |
| "grad_norm": 4.71875, | |
| "learning_rate": 4.3206389875579886e-05, | |
| "loss": 1.4496, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.43673207942700754, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 4.272113200954988e-05, | |
| "loss": 1.4421, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.465847551388808, | |
| "grad_norm": 4.90625, | |
| "learning_rate": 4.223587414351987e-05, | |
| "loss": 1.4305, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.4949630233506085, | |
| "grad_norm": 4.25, | |
| "learning_rate": 4.175061627748986e-05, | |
| "loss": 1.4251, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.524078495312409, | |
| "grad_norm": 4.53125, | |
| "learning_rate": 4.126535841145985e-05, | |
| "loss": 1.4116, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.5531939672742096, | |
| "grad_norm": 4.375, | |
| "learning_rate": 4.0780100545429844e-05, | |
| "loss": 1.4124, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.58230943923601, | |
| "grad_norm": 5.40625, | |
| "learning_rate": 4.0294842679399835e-05, | |
| "loss": 1.4035, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.58230943923601, | |
| "eval_accuracy": 0.7037869026306487, | |
| "eval_loss": 1.3742600679397583, | |
| "eval_runtime": 6.4605, | |
| "eval_samples_per_second": 89.931, | |
| "eval_steps_per_second": 2.941, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.6114249111978105, | |
| "grad_norm": 4.40625, | |
| "learning_rate": 3.980958481336983e-05, | |
| "loss": 1.4042, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 0.640540383159611, | |
| "grad_norm": 4.96875, | |
| "learning_rate": 3.932432694733982e-05, | |
| "loss": 1.3966, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.6696558551214116, | |
| "grad_norm": 4.5625, | |
| "learning_rate": 3.883906908130981e-05, | |
| "loss": 1.3887, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 0.698771327083212, | |
| "grad_norm": 5.0, | |
| "learning_rate": 3.83538112152798e-05, | |
| "loss": 1.3843, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.7278867990450125, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 3.7868553349249794e-05, | |
| "loss": 1.3783, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.757002271006813, | |
| "grad_norm": 4.6875, | |
| "learning_rate": 3.7383295483219785e-05, | |
| "loss": 1.375, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.7861177429686135, | |
| "grad_norm": 4.46875, | |
| "learning_rate": 3.689803761718978e-05, | |
| "loss": 1.3756, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 0.815233214930414, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 3.641277975115977e-05, | |
| "loss": 1.3636, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.8443486868922145, | |
| "grad_norm": 4.5625, | |
| "learning_rate": 3.592752188512976e-05, | |
| "loss": 1.3688, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 0.8734641588540151, | |
| "grad_norm": 4.59375, | |
| "learning_rate": 3.544226401909975e-05, | |
| "loss": 1.3652, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.9025796308158155, | |
| "grad_norm": 4.90625, | |
| "learning_rate": 3.4957006153069744e-05, | |
| "loss": 1.3611, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 0.931695102777616, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 3.4471748287039735e-05, | |
| "loss": 1.352, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.9608105747394166, | |
| "grad_norm": 4.625, | |
| "learning_rate": 3.398649042100973e-05, | |
| "loss": 1.3562, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 0.989926046701217, | |
| "grad_norm": 4.84375, | |
| "learning_rate": 3.350123255497972e-05, | |
| "loss": 1.3491, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.0190415186630175, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 3.301597468894971e-05, | |
| "loss": 1.3448, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.048156990624818, | |
| "grad_norm": 4.625, | |
| "learning_rate": 3.25307168229197e-05, | |
| "loss": 1.344, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.0772724625866186, | |
| "grad_norm": 4.75, | |
| "learning_rate": 3.204545895688969e-05, | |
| "loss": 1.3381, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 1.106387934548419, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 3.1560201090859685e-05, | |
| "loss": 1.34, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.1355034065102196, | |
| "grad_norm": 5.375, | |
| "learning_rate": 3.107494322482968e-05, | |
| "loss": 1.3352, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 1.16461887847202, | |
| "grad_norm": 5.71875, | |
| "learning_rate": 3.058968535879967e-05, | |
| "loss": 1.3374, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.16461887847202, | |
| "eval_accuracy": 0.7118764765058947, | |
| "eval_loss": 1.3248625993728638, | |
| "eval_runtime": 6.4569, | |
| "eval_samples_per_second": 89.981, | |
| "eval_steps_per_second": 2.943, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.1937343504338205, | |
| "grad_norm": 4.5, | |
| "learning_rate": 3.010442749276966e-05, | |
| "loss": 1.3314, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 1.222849822395621, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 2.9619169626739652e-05, | |
| "loss": 1.3296, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 1.2519652943574215, | |
| "grad_norm": 4.40625, | |
| "learning_rate": 2.9133911760709644e-05, | |
| "loss": 1.3331, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 1.281080766319222, | |
| "grad_norm": 4.75, | |
| "learning_rate": 2.8648653894679632e-05, | |
| "loss": 1.332, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 1.3101962382810226, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 2.8163396028649623e-05, | |
| "loss": 1.3291, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 1.3393117102428231, | |
| "grad_norm": 4.53125, | |
| "learning_rate": 2.7678138162619615e-05, | |
| "loss": 1.3301, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 1.3684271822046234, | |
| "grad_norm": 4.8125, | |
| "learning_rate": 2.7192880296589607e-05, | |
| "loss": 1.3266, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 1.397542654166424, | |
| "grad_norm": 5.0, | |
| "learning_rate": 2.67076224305596e-05, | |
| "loss": 1.3178, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 1.4266581261282245, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 2.6222364564529593e-05, | |
| "loss": 1.323, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 1.455773598090025, | |
| "grad_norm": 4.78125, | |
| "learning_rate": 2.5737106698499585e-05, | |
| "loss": 1.3206, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.4848890700518256, | |
| "grad_norm": 5.0625, | |
| "learning_rate": 2.5251848832469577e-05, | |
| "loss": 1.319, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 1.5140045420136259, | |
| "grad_norm": 5.34375, | |
| "learning_rate": 2.476659096643957e-05, | |
| "loss": 1.3129, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 1.5431200139754266, | |
| "grad_norm": 4.96875, | |
| "learning_rate": 2.428133310040956e-05, | |
| "loss": 1.3163, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 1.572235485937227, | |
| "grad_norm": 4.59375, | |
| "learning_rate": 2.3796075234379552e-05, | |
| "loss": 1.313, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 1.6013509578990277, | |
| "grad_norm": 4.40625, | |
| "learning_rate": 2.331081736834954e-05, | |
| "loss": 1.3167, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 1.630466429860828, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 2.282555950231953e-05, | |
| "loss": 1.3151, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 1.6595819018226285, | |
| "grad_norm": 5.09375, | |
| "learning_rate": 2.2340301636289527e-05, | |
| "loss": 1.312, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 1.688697373784429, | |
| "grad_norm": 4.71875, | |
| "learning_rate": 2.185504377025952e-05, | |
| "loss": 1.3196, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 1.7178128457462294, | |
| "grad_norm": 4.46875, | |
| "learning_rate": 2.136978590422951e-05, | |
| "loss": 1.3149, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 1.7469283177080301, | |
| "grad_norm": 4.71875, | |
| "learning_rate": 2.0884528038199498e-05, | |
| "loss": 1.3142, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 1.7469283177080301, | |
| "eval_accuracy": 0.7175331317769543, | |
| "eval_loss": 1.2938237190246582, | |
| "eval_runtime": 6.453, | |
| "eval_samples_per_second": 90.036, | |
| "eval_steps_per_second": 2.944, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 1.7760437896698305, | |
| "grad_norm": 5.03125, | |
| "learning_rate": 2.039927017216949e-05, | |
| "loss": 1.3115, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 1.8051592616316312, | |
| "grad_norm": 4.90625, | |
| "learning_rate": 1.9914012306139485e-05, | |
| "loss": 1.3155, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 1.8342747335934315, | |
| "grad_norm": 5.0, | |
| "learning_rate": 1.9428754440109477e-05, | |
| "loss": 1.3179, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 1.863390205555232, | |
| "grad_norm": 5.21875, | |
| "learning_rate": 1.8943496574079468e-05, | |
| "loss": 1.3091, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 1.8925056775170326, | |
| "grad_norm": 4.46875, | |
| "learning_rate": 1.8458238708049457e-05, | |
| "loss": 1.3148, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 1.921621149478833, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 1.7972980842019448e-05, | |
| "loss": 1.3091, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 1.9507366214406336, | |
| "grad_norm": 5.03125, | |
| "learning_rate": 1.7487722975989443e-05, | |
| "loss": 1.3176, | |
| "step": 33500 | |
| }, | |
| { | |
| "epoch": 1.979852093402434, | |
| "grad_norm": 4.625, | |
| "learning_rate": 1.7002465109959435e-05, | |
| "loss": 1.3135, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 2.0089675653642347, | |
| "grad_norm": 5.03125, | |
| "learning_rate": 1.6517207243929427e-05, | |
| "loss": 1.3134, | |
| "step": 34500 | |
| }, | |
| { | |
| "epoch": 2.038083037326035, | |
| "grad_norm": 4.875, | |
| "learning_rate": 1.6031949377899415e-05, | |
| "loss": 1.3061, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 2.0671985092878353, | |
| "grad_norm": 4.84375, | |
| "learning_rate": 1.5546691511869406e-05, | |
| "loss": 1.3053, | |
| "step": 35500 | |
| }, | |
| { | |
| "epoch": 2.096313981249636, | |
| "grad_norm": 4.40625, | |
| "learning_rate": 1.5061433645839398e-05, | |
| "loss": 1.3102, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 2.1254294532114364, | |
| "grad_norm": 4.90625, | |
| "learning_rate": 1.4576175779809393e-05, | |
| "loss": 1.3137, | |
| "step": 36500 | |
| }, | |
| { | |
| "epoch": 2.154544925173237, | |
| "grad_norm": 4.75, | |
| "learning_rate": 1.4090917913779383e-05, | |
| "loss": 1.306, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 2.1836603971350375, | |
| "grad_norm": 4.9375, | |
| "learning_rate": 1.3605660047749375e-05, | |
| "loss": 1.3029, | |
| "step": 37500 | |
| }, | |
| { | |
| "epoch": 2.212775869096838, | |
| "grad_norm": 5.34375, | |
| "learning_rate": 1.3120402181719366e-05, | |
| "loss": 1.3068, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 2.2418913410586385, | |
| "grad_norm": 4.96875, | |
| "learning_rate": 1.2635144315689356e-05, | |
| "loss": 1.3033, | |
| "step": 38500 | |
| }, | |
| { | |
| "epoch": 2.2710068130204393, | |
| "grad_norm": 5.46875, | |
| "learning_rate": 1.214988644965935e-05, | |
| "loss": 1.3104, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 2.3001222849822396, | |
| "grad_norm": 4.71875, | |
| "learning_rate": 1.1664628583629341e-05, | |
| "loss": 1.3058, | |
| "step": 39500 | |
| }, | |
| { | |
| "epoch": 2.32923775694404, | |
| "grad_norm": 4.53125, | |
| "learning_rate": 1.1179370717599333e-05, | |
| "loss": 1.3064, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.32923775694404, | |
| "eval_accuracy": 0.7175919626253005, | |
| "eval_loss": 1.2843679189682007, | |
| "eval_runtime": 12.4622, | |
| "eval_samples_per_second": 46.621, | |
| "eval_steps_per_second": 1.525, | |
| "step": 40000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 51519, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 10000, | |
| "total_flos": 2.5163609407488e+18, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |