|
{ |
|
"best_metric": 0.5380504727363586, |
|
"best_model_checkpoint": "/data/jcanete/all_results/xnli/beto_uncased/epochs_4_bs_64_lr_5e-5/checkpoint-12000", |
|
"epoch": 4.0, |
|
"global_step": 24544, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.08, |
|
"learning_rate": 4.898142112125163e-05, |
|
"loss": 0.7866, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 0.16, |
|
"learning_rate": 4.796487940026076e-05, |
|
"loss": 0.703, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 0.24, |
|
"learning_rate": 4.6946300521512387e-05, |
|
"loss": 0.6809, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 0.33, |
|
"learning_rate": 4.5927721642764015e-05, |
|
"loss": 0.6555, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 0.33, |
|
"eval_accuracy": 0.7497991967871486, |
|
"eval_loss": 0.602264940738678, |
|
"eval_runtime": 1.3367, |
|
"eval_samples_per_second": 1862.863, |
|
"eval_steps_per_second": 29.177, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 0.41, |
|
"learning_rate": 4.490914276401565e-05, |
|
"loss": 0.6351, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 0.49, |
|
"learning_rate": 4.389056388526728e-05, |
|
"loss": 0.6267, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 0.57, |
|
"learning_rate": 4.2871985006518906e-05, |
|
"loss": 0.62, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 0.65, |
|
"learning_rate": 4.185340612777054e-05, |
|
"loss": 0.6094, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 0.65, |
|
"eval_accuracy": 0.7678714859437751, |
|
"eval_loss": 0.5793344378471375, |
|
"eval_runtime": 2.4971, |
|
"eval_samples_per_second": 997.156, |
|
"eval_steps_per_second": 15.618, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 0.73, |
|
"learning_rate": 4.083482724902217e-05, |
|
"loss": 0.5983, |
|
"step": 4500 |
|
}, |
|
{ |
|
"epoch": 0.81, |
|
"learning_rate": 3.9816248370273797e-05, |
|
"loss": 0.5886, |
|
"step": 5000 |
|
}, |
|
{ |
|
"epoch": 0.9, |
|
"learning_rate": 3.880174380704042e-05, |
|
"loss": 0.5871, |
|
"step": 5500 |
|
}, |
|
{ |
|
"epoch": 0.98, |
|
"learning_rate": 3.7783164928292046e-05, |
|
"loss": 0.5771, |
|
"step": 6000 |
|
}, |
|
{ |
|
"epoch": 0.98, |
|
"eval_accuracy": 0.7811244979919679, |
|
"eval_loss": 0.5455663800239563, |
|
"eval_runtime": 2.4323, |
|
"eval_samples_per_second": 1023.703, |
|
"eval_steps_per_second": 16.034, |
|
"step": 6000 |
|
}, |
|
{ |
|
"epoch": 1.06, |
|
"learning_rate": 3.676458604954368e-05, |
|
"loss": 0.4853, |
|
"step": 6500 |
|
}, |
|
{ |
|
"epoch": 1.14, |
|
"learning_rate": 3.574600717079531e-05, |
|
"loss": 0.4519, |
|
"step": 7000 |
|
}, |
|
{ |
|
"epoch": 1.22, |
|
"learning_rate": 3.472742829204694e-05, |
|
"loss": 0.4597, |
|
"step": 7500 |
|
}, |
|
{ |
|
"epoch": 1.3, |
|
"learning_rate": 3.3708849413298565e-05, |
|
"loss": 0.4484, |
|
"step": 8000 |
|
}, |
|
{ |
|
"epoch": 1.3, |
|
"eval_accuracy": 0.7899598393574297, |
|
"eval_loss": 0.5768685340881348, |
|
"eval_runtime": 1.3402, |
|
"eval_samples_per_second": 1857.904, |
|
"eval_steps_per_second": 29.1, |
|
"step": 8000 |
|
}, |
|
{ |
|
"epoch": 1.39, |
|
"learning_rate": 3.26902705345502e-05, |
|
"loss": 0.4567, |
|
"step": 8500 |
|
}, |
|
{ |
|
"epoch": 1.47, |
|
"learning_rate": 3.167372881355932e-05, |
|
"loss": 0.4483, |
|
"step": 9000 |
|
}, |
|
{ |
|
"epoch": 1.55, |
|
"learning_rate": 3.065514993481095e-05, |
|
"loss": 0.446, |
|
"step": 9500 |
|
}, |
|
{ |
|
"epoch": 1.63, |
|
"learning_rate": 2.9636571056062584e-05, |
|
"loss": 0.4513, |
|
"step": 10000 |
|
}, |
|
{ |
|
"epoch": 1.63, |
|
"eval_accuracy": 0.7738955823293173, |
|
"eval_loss": 0.5736198425292969, |
|
"eval_runtime": 1.3279, |
|
"eval_samples_per_second": 1875.149, |
|
"eval_steps_per_second": 29.37, |
|
"step": 10000 |
|
}, |
|
{ |
|
"epoch": 1.71, |
|
"learning_rate": 2.8617992177314212e-05, |
|
"loss": 0.4511, |
|
"step": 10500 |
|
}, |
|
{ |
|
"epoch": 1.79, |
|
"learning_rate": 2.759941329856584e-05, |
|
"loss": 0.4423, |
|
"step": 11000 |
|
}, |
|
{ |
|
"epoch": 1.87, |
|
"learning_rate": 2.6580834419817475e-05, |
|
"loss": 0.4472, |
|
"step": 11500 |
|
}, |
|
{ |
|
"epoch": 1.96, |
|
"learning_rate": 2.5564292698826596e-05, |
|
"loss": 0.4362, |
|
"step": 12000 |
|
}, |
|
{ |
|
"epoch": 1.96, |
|
"eval_accuracy": 0.7811244979919679, |
|
"eval_loss": 0.5380504727363586, |
|
"eval_runtime": 1.33, |
|
"eval_samples_per_second": 1872.162, |
|
"eval_steps_per_second": 29.323, |
|
"step": 12000 |
|
}, |
|
{ |
|
"epoch": 2.04, |
|
"learning_rate": 2.4545713820078228e-05, |
|
"loss": 0.3786, |
|
"step": 12500 |
|
}, |
|
{ |
|
"epoch": 2.12, |
|
"learning_rate": 2.3527134941329856e-05, |
|
"loss": 0.2879, |
|
"step": 13000 |
|
}, |
|
{ |
|
"epoch": 2.2, |
|
"learning_rate": 2.2508556062581487e-05, |
|
"loss": 0.2885, |
|
"step": 13500 |
|
}, |
|
{ |
|
"epoch": 2.28, |
|
"learning_rate": 2.148997718383312e-05, |
|
"loss": 0.2909, |
|
"step": 14000 |
|
}, |
|
{ |
|
"epoch": 2.28, |
|
"eval_accuracy": 0.7835341365461848, |
|
"eval_loss": 0.6650500297546387, |
|
"eval_runtime": 2.4288, |
|
"eval_samples_per_second": 1025.198, |
|
"eval_steps_per_second": 16.057, |
|
"step": 14000 |
|
}, |
|
{ |
|
"epoch": 2.36, |
|
"learning_rate": 2.0473435462842243e-05, |
|
"loss": 0.2991, |
|
"step": 14500 |
|
}, |
|
{ |
|
"epoch": 2.44, |
|
"learning_rate": 1.945485658409387e-05, |
|
"loss": 0.2865, |
|
"step": 15000 |
|
}, |
|
{ |
|
"epoch": 2.53, |
|
"learning_rate": 1.8436277705345503e-05, |
|
"loss": 0.2947, |
|
"step": 15500 |
|
}, |
|
{ |
|
"epoch": 2.61, |
|
"learning_rate": 1.741769882659713e-05, |
|
"loss": 0.3002, |
|
"step": 16000 |
|
}, |
|
{ |
|
"epoch": 2.61, |
|
"eval_accuracy": 0.7807228915662651, |
|
"eval_loss": 0.6428889632225037, |
|
"eval_runtime": 2.4279, |
|
"eval_samples_per_second": 1025.584, |
|
"eval_steps_per_second": 16.063, |
|
"step": 16000 |
|
}, |
|
{ |
|
"epoch": 2.69, |
|
"learning_rate": 1.6399119947848762e-05, |
|
"loss": 0.2916, |
|
"step": 16500 |
|
}, |
|
{ |
|
"epoch": 2.77, |
|
"learning_rate": 1.5380541069100394e-05, |
|
"loss": 0.2987, |
|
"step": 17000 |
|
}, |
|
{ |
|
"epoch": 2.85, |
|
"learning_rate": 1.4361962190352022e-05, |
|
"loss": 0.3005, |
|
"step": 17500 |
|
}, |
|
{ |
|
"epoch": 2.93, |
|
"learning_rate": 1.3343383311603652e-05, |
|
"loss": 0.2916, |
|
"step": 18000 |
|
}, |
|
{ |
|
"epoch": 2.93, |
|
"eval_accuracy": 0.7779116465863454, |
|
"eval_loss": 0.6353530883789062, |
|
"eval_runtime": 1.3158, |
|
"eval_samples_per_second": 1892.365, |
|
"eval_steps_per_second": 29.639, |
|
"step": 18000 |
|
}, |
|
{ |
|
"epoch": 3.01, |
|
"learning_rate": 1.2328878748370275e-05, |
|
"loss": 0.2756, |
|
"step": 18500 |
|
}, |
|
{ |
|
"epoch": 3.1, |
|
"learning_rate": 1.1310299869621905e-05, |
|
"loss": 0.1773, |
|
"step": 19000 |
|
}, |
|
{ |
|
"epoch": 3.18, |
|
"learning_rate": 1.0291720990873534e-05, |
|
"loss": 0.1788, |
|
"step": 19500 |
|
}, |
|
{ |
|
"epoch": 3.26, |
|
"learning_rate": 9.273142112125164e-06, |
|
"loss": 0.1801, |
|
"step": 20000 |
|
}, |
|
{ |
|
"epoch": 3.26, |
|
"eval_accuracy": 0.7811244979919679, |
|
"eval_loss": 0.7538436055183411, |
|
"eval_runtime": 1.3281, |
|
"eval_samples_per_second": 1874.884, |
|
"eval_steps_per_second": 29.366, |
|
"step": 20000 |
|
}, |
|
{ |
|
"epoch": 3.34, |
|
"learning_rate": 8.254563233376792e-06, |
|
"loss": 0.176, |
|
"step": 20500 |
|
}, |
|
{ |
|
"epoch": 3.42, |
|
"learning_rate": 7.23802151238592e-06, |
|
"loss": 0.1801, |
|
"step": 21000 |
|
}, |
|
{ |
|
"epoch": 3.5, |
|
"learning_rate": 6.219442633637549e-06, |
|
"loss": 0.1761, |
|
"step": 21500 |
|
}, |
|
{ |
|
"epoch": 3.59, |
|
"learning_rate": 5.200863754889179e-06, |
|
"loss": 0.1795, |
|
"step": 22000 |
|
}, |
|
{ |
|
"epoch": 3.59, |
|
"eval_accuracy": 0.7891566265060241, |
|
"eval_loss": 0.7428993582725525, |
|
"eval_runtime": 2.4875, |
|
"eval_samples_per_second": 1000.986, |
|
"eval_steps_per_second": 15.678, |
|
"step": 22000 |
|
}, |
|
{ |
|
"epoch": 3.67, |
|
"learning_rate": 4.182284876140809e-06, |
|
"loss": 0.1734, |
|
"step": 22500 |
|
}, |
|
{ |
|
"epoch": 3.75, |
|
"learning_rate": 3.1637059973924383e-06, |
|
"loss": 0.1759, |
|
"step": 23000 |
|
}, |
|
{ |
|
"epoch": 3.83, |
|
"learning_rate": 2.1471642764015647e-06, |
|
"loss": 0.1789, |
|
"step": 23500 |
|
}, |
|
{ |
|
"epoch": 3.91, |
|
"learning_rate": 1.1285853976531943e-06, |
|
"loss": 0.1723, |
|
"step": 24000 |
|
}, |
|
{ |
|
"epoch": 3.91, |
|
"eval_accuracy": 0.7895582329317269, |
|
"eval_loss": 0.7440813183784485, |
|
"eval_runtime": 2.4411, |
|
"eval_samples_per_second": 1020.029, |
|
"eval_steps_per_second": 15.976, |
|
"step": 24000 |
|
}, |
|
{ |
|
"epoch": 3.99, |
|
"learning_rate": 1.1000651890482399e-07, |
|
"loss": 0.17, |
|
"step": 24500 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"step": 24544, |
|
"total_flos": 8.969723878657651e+16, |
|
"train_loss": 0.38933088829573315, |
|
"train_runtime": 9866.0231, |
|
"train_samples_per_second": 159.214, |
|
"train_steps_per_second": 2.488 |
|
} |
|
], |
|
"max_steps": 24544, |
|
"num_train_epochs": 4, |
|
"total_flos": 8.969723878657651e+16, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|