| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "global_step": 42311, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 5.907372400756144e-06, | |
| "loss": 1.9951, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.1814744801512288e-05, | |
| "loss": 1.4127, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1.7722117202268433e-05, | |
| "loss": 1.3188, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.3629489603024576e-05, | |
| "loss": 1.2728, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 2.953686200378072e-05, | |
| "loss": 1.248, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 3.5444234404536866e-05, | |
| "loss": 1.2485, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.135160680529301e-05, | |
| "loss": 1.2476, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.725897920604915e-05, | |
| "loss": 1.259, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.9648100002626124e-05, | |
| "loss": 1.2954, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.899157015677933e-05, | |
| "loss": 1.3149, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.833504031093254e-05, | |
| "loss": 1.3251, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.767851046508575e-05, | |
| "loss": 1.342, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.702198061923895e-05, | |
| "loss": 1.344, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.636545077339216e-05, | |
| "loss": 1.335, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.570892092754537e-05, | |
| "loss": 1.3276, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.5052391081698575e-05, | |
| "loss": 1.3311, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.4395861235851784e-05, | |
| "loss": 1.3312, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.373933139000499e-05, | |
| "loss": 1.3298, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.30828015441582e-05, | |
| "loss": 1.3236, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.242627169831141e-05, | |
| "loss": 1.3251, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.176974185246462e-05, | |
| "loss": 1.3347, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.1113212006617825e-05, | |
| "loss": 1.3309, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.0456682160771034e-05, | |
| "loss": 1.3392, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 3.9800152314924235e-05, | |
| "loss": 1.3407, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 3.914362246907744e-05, | |
| "loss": 1.3494, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 3.848709262323065e-05, | |
| "loss": 1.3534, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 3.783056277738386e-05, | |
| "loss": 1.3561, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 3.717403293153707e-05, | |
| "loss": 1.3614, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 3.651750308569028e-05, | |
| "loss": 1.3675, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 3.5860973239843485e-05, | |
| "loss": 1.3728, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.520444339399669e-05, | |
| "loss": 1.3832, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.45479135481499e-05, | |
| "loss": 1.3866, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.389138370230311e-05, | |
| "loss": 1.4043, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.323485385645632e-05, | |
| "loss": 1.4023, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.2578324010609527e-05, | |
| "loss": 1.4125, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.192179416476273e-05, | |
| "loss": 1.4165, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.1265264318915936e-05, | |
| "loss": 1.4195, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.0608734473069145e-05, | |
| "loss": 1.4239, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 2.9952204627222353e-05, | |
| "loss": 1.4311, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 2.9295674781375565e-05, | |
| "loss": 1.4391, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 2.8639144935528773e-05, | |
| "loss": 1.4343, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.7982615089681978e-05, | |
| "loss": 1.4432, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.7326085243835186e-05, | |
| "loss": 1.4515, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.6669555397988395e-05, | |
| "loss": 1.4579, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.60130255521416e-05, | |
| "loss": 1.4601, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.5356495706294808e-05, | |
| "loss": 1.4647, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.469996586044802e-05, | |
| "loss": 1.4735, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.4043436014601224e-05, | |
| "loss": 1.4769, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.3386906168754433e-05, | |
| "loss": 1.4876, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.273037632290764e-05, | |
| "loss": 1.49, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.2073846477060846e-05, | |
| "loss": 1.4942, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 2.1417316631214058e-05, | |
| "loss": 1.4988, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 2.0760786785367266e-05, | |
| "loss": 1.5122, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 2.010425693952047e-05, | |
| "loss": 1.5134, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.944772709367368e-05, | |
| "loss": 1.5191, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.8791197247826888e-05, | |
| "loss": 1.5193, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.8134667401980096e-05, | |
| "loss": 1.5262, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.7478137556133304e-05, | |
| "loss": 1.5383, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.682160771028651e-05, | |
| "loss": 1.5349, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.6165077864439717e-05, | |
| "loss": 1.5411, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.5508548018592926e-05, | |
| "loss": 1.5512, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.4852018172746136e-05, | |
| "loss": 1.5535, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 1.4195488326899342e-05, | |
| "loss": 1.5574, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 1.3538958481052549e-05, | |
| "loss": 1.5595, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 1.2882428635205757e-05, | |
| "loss": 1.5647, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 1.2225898789358966e-05, | |
| "loss": 1.566, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 1.1569368943512172e-05, | |
| "loss": 1.5691, | |
| "step": 33500 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 1.091283909766538e-05, | |
| "loss": 1.5736, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 1.0256309251818589e-05, | |
| "loss": 1.5762, | |
| "step": 34500 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 9.599779405971796e-06, | |
| "loss": 1.5778, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 8.943249560125004e-06, | |
| "loss": 1.5796, | |
| "step": 35500 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 8.286719714278212e-06, | |
| "loss": 1.574, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 7.63018986843142e-06, | |
| "loss": 1.5776, | |
| "step": 36500 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 6.973660022584627e-06, | |
| "loss": 1.5834, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 6.3171301767378345e-06, | |
| "loss": 1.5949, | |
| "step": 37500 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 5.660600330891042e-06, | |
| "loss": 1.582, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 5.00407048504425e-06, | |
| "loss": 1.5904, | |
| "step": 38500 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 4.347540639197458e-06, | |
| "loss": 1.595, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 3.691010793350666e-06, | |
| "loss": 1.5847, | |
| "step": 39500 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 3.0344809475038736e-06, | |
| "loss": 1.5795, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.3779511016570815e-06, | |
| "loss": 1.5827, | |
| "step": 40500 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 1.7214212558102892e-06, | |
| "loss": 1.5895, | |
| "step": 41000 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 1.064891409963497e-06, | |
| "loss": 1.5856, | |
| "step": 41500 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.083615641167047e-07, | |
| "loss": 1.5763, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 42311, | |
| "total_flos": 4.127720706618163e+17, | |
| "train_loss": 1.455913220122233, | |
| "train_runtime": 36443.3148, | |
| "train_samples_per_second": 37.152, | |
| "train_steps_per_second": 1.161 | |
| } | |
| ], | |
| "max_steps": 42311, | |
| "num_train_epochs": 1, | |
| "total_flos": 4.127720706618163e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |