| { | |
| "best_metric": 0.08294033469935763, | |
| "best_model_checkpoint": "/bartabsa-reproduce/outputs/gpt22gpt2_42_std_paper/checkpoint-8000", | |
| "epoch": 2.999832822513235, | |
| "eval_steps": 2000, | |
| "global_step": 26916, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05572582892170521, | |
| "grad_norm": 2.4936771392822266, | |
| "learning_rate": 2.5e-05, | |
| "loss": 3.555, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.11145165784341042, | |
| "grad_norm": 1.7875964641571045, | |
| "learning_rate": 5e-05, | |
| "loss": 2.9668, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.16717748676511562, | |
| "grad_norm": 1.4690954685211182, | |
| "learning_rate": 4.9035344960642076e-05, | |
| "loss": 2.8594, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.22290331568682084, | |
| "grad_norm": 1.5910253524780273, | |
| "learning_rate": 4.807068992128415e-05, | |
| "loss": 2.7391, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.22290331568682084, | |
| "eval_loss": 2.587376117706299, | |
| "eval_rouge1": 0.15890436461961194, | |
| "eval_rouge2": 0.026691787346904763, | |
| "eval_rougeL": 0.11299625068223401, | |
| "eval_rougeLsum": 0.15080434380335367, | |
| "eval_runtime": 6206.3629, | |
| "eval_samples_per_second": 2.154, | |
| "eval_steps_per_second": 0.135, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.27862914460852606, | |
| "grad_norm": 1.6178048849105835, | |
| "learning_rate": 4.7106034881926225e-05, | |
| "loss": 2.6739, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.33435497353023125, | |
| "grad_norm": 1.9126936197280884, | |
| "learning_rate": 4.61413798425683e-05, | |
| "loss": 2.6193, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.3900808024519365, | |
| "grad_norm": 1.5849372148513794, | |
| "learning_rate": 4.517672480321037e-05, | |
| "loss": 2.5813, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.4458066313736417, | |
| "grad_norm": 5.0478129386901855, | |
| "learning_rate": 4.421206976385245e-05, | |
| "loss": 2.5412, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.4458066313736417, | |
| "eval_loss": 2.4282822608947754, | |
| "eval_rouge1": 0.26692339384492103, | |
| "eval_rouge2": 0.06788970022541349, | |
| "eval_rougeL": 0.16057716127224064, | |
| "eval_rougeLsum": 0.24985170034027104, | |
| "eval_runtime": 6305.6599, | |
| "eval_samples_per_second": 2.12, | |
| "eval_steps_per_second": 0.133, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.5015324602953469, | |
| "grad_norm": 1.589734673500061, | |
| "learning_rate": 4.324741472449452e-05, | |
| "loss": 2.5082, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.5572582892170521, | |
| "grad_norm": 1.5139461755752563, | |
| "learning_rate": 4.2282759685136595e-05, | |
| "loss": 2.4859, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.6129841181387573, | |
| "grad_norm": 1.4664047956466675, | |
| "learning_rate": 4.131810464577867e-05, | |
| "loss": 2.4547, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.6687099470604625, | |
| "grad_norm": 1.7493504285812378, | |
| "learning_rate": 4.035344960642074e-05, | |
| "loss": 2.4358, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.6687099470604625, | |
| "eval_loss": 2.345005512237549, | |
| "eval_rouge1": 0.2730819963251183, | |
| "eval_rouge2": 0.07305508048248258, | |
| "eval_rougeL": 0.1637512659125704, | |
| "eval_rougeLsum": 0.25593027675446556, | |
| "eval_runtime": 6286.1193, | |
| "eval_samples_per_second": 2.127, | |
| "eval_steps_per_second": 0.133, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.7244357759821677, | |
| "grad_norm": 1.757111668586731, | |
| "learning_rate": 3.938879456706282e-05, | |
| "loss": 2.4158, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.780161604903873, | |
| "grad_norm": 1.46837317943573, | |
| "learning_rate": 3.84241395277049e-05, | |
| "loss": 2.3926, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.8358874338255782, | |
| "grad_norm": 1.4906083345413208, | |
| "learning_rate": 3.745948448834697e-05, | |
| "loss": 2.3803, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.8916132627472834, | |
| "grad_norm": 1.6719425916671753, | |
| "learning_rate": 3.6494829448989046e-05, | |
| "loss": 2.3592, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.8916132627472834, | |
| "eval_loss": 2.261608600616455, | |
| "eval_rouge1": 0.29601660868967444, | |
| "eval_rouge2": 0.08294033469935763, | |
| "eval_rougeL": 0.17290133105856215, | |
| "eval_rougeLsum": 0.2765445625043713, | |
| "eval_runtime": 6302.0707, | |
| "eval_samples_per_second": 2.121, | |
| "eval_steps_per_second": 0.133, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.9473390916689886, | |
| "grad_norm": 1.585368275642395, | |
| "learning_rate": 3.553017440963112e-05, | |
| "loss": 2.3381, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.0030649205906939, | |
| "grad_norm": NaN, | |
| "learning_rate": 3.4565519370273194e-05, | |
| "loss": 2.7416, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.058790749512399, | |
| "grad_norm": NaN, | |
| "learning_rate": 3.360086433091527e-05, | |
| "loss": 0.0, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.1145165784341042, | |
| "grad_norm": NaN, | |
| "learning_rate": 3.263620929155734e-05, | |
| "loss": 0.0, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.1145165784341042, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5771.2576, | |
| "eval_samples_per_second": 2.316, | |
| "eval_steps_per_second": 0.145, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.1702424073558095, | |
| "grad_norm": NaN, | |
| "learning_rate": 3.1671554252199416e-05, | |
| "loss": 0.0, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.2259682362775146, | |
| "grad_norm": NaN, | |
| "learning_rate": 3.070689921284149e-05, | |
| "loss": 0.0, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.28169406519922, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.9742244173483564e-05, | |
| "loss": 0.0, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.337419894120925, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.8777589134125638e-05, | |
| "loss": 0.0, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.337419894120925, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5768.1931, | |
| "eval_samples_per_second": 2.318, | |
| "eval_steps_per_second": 0.145, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.3931457230426303, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.7812934094767712e-05, | |
| "loss": 0.0, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.4488715519643356, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.6848279055409786e-05, | |
| "loss": 0.0, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.5045973808860407, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.588362401605186e-05, | |
| "loss": 0.0, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.5603232098077457, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.4918968976693934e-05, | |
| "loss": 0.0, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.5603232098077457, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5772.8437, | |
| "eval_samples_per_second": 2.316, | |
| "eval_steps_per_second": 0.145, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.616049038729451, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.3954313937336008e-05, | |
| "loss": 0.0, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.6717748676511563, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.2989658897978082e-05, | |
| "loss": 0.0, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.7275006965728616, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.2025003858620156e-05, | |
| "loss": 0.0, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.7832265254945667, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.1060348819262234e-05, | |
| "loss": 0.0, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.7832265254945667, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5773.9768, | |
| "eval_samples_per_second": 2.315, | |
| "eval_steps_per_second": 0.145, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.8389523544162718, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.0095693779904308e-05, | |
| "loss": 0.0, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.894678183337977, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.9131038740546382e-05, | |
| "loss": 0.0, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.9504040122596824, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.8166383701188456e-05, | |
| "loss": 0.0, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.0061298411813877, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.720172866183053e-05, | |
| "loss": 0.0, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.0061298411813877, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5771.1419, | |
| "eval_samples_per_second": 2.316, | |
| "eval_steps_per_second": 0.145, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.0618556701030926, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.6237073622472604e-05, | |
| "loss": 0.0, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.117581499024798, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.5272418583114678e-05, | |
| "loss": 0.0, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.173307327946503, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.4307763543756752e-05, | |
| "loss": 0.0, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.2290331568682085, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.3343108504398828e-05, | |
| "loss": 0.0, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.2290331568682085, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5768.4551, | |
| "eval_samples_per_second": 2.317, | |
| "eval_steps_per_second": 0.145, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.2847589857899138, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.2378453465040902e-05, | |
| "loss": 0.0, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.340484814711619, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.1413798425682977e-05, | |
| "loss": 0.0, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.396210643633324, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.0449143386325052e-05, | |
| "loss": 0.0, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 2.4519364725550292, | |
| "grad_norm": NaN, | |
| "learning_rate": 9.484488346967126e-06, | |
| "loss": 0.0, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.4519364725550292, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5770.7442, | |
| "eval_samples_per_second": 2.317, | |
| "eval_steps_per_second": 0.145, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.5076623014767345, | |
| "grad_norm": NaN, | |
| "learning_rate": 8.5198333076092e-06, | |
| "loss": 0.0, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.56338813039844, | |
| "grad_norm": NaN, | |
| "learning_rate": 7.5551782682512745e-06, | |
| "loss": 0.0, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 2.6191139593201447, | |
| "grad_norm": NaN, | |
| "learning_rate": 6.5905232288933485e-06, | |
| "loss": 0.0, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 2.67483978824185, | |
| "grad_norm": NaN, | |
| "learning_rate": 5.6258681895354226e-06, | |
| "loss": 0.0, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.67483978824185, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5767.1561, | |
| "eval_samples_per_second": 2.318, | |
| "eval_steps_per_second": 0.145, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.7305656171635553, | |
| "grad_norm": NaN, | |
| "learning_rate": 4.661213150177497e-06, | |
| "loss": 0.0, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 2.7862914460852606, | |
| "grad_norm": NaN, | |
| "learning_rate": 3.6965581108195706e-06, | |
| "loss": 0.0, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.842017275006966, | |
| "grad_norm": NaN, | |
| "learning_rate": 2.7319030714616455e-06, | |
| "loss": 0.0, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 2.897743103928671, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.7672480321037198e-06, | |
| "loss": 0.0, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 2.897743103928671, | |
| "eval_loss": NaN, | |
| "eval_rouge1": 0.0, | |
| "eval_rouge2": 0.0, | |
| "eval_rougeL": 0.0, | |
| "eval_rougeLsum": 0.0, | |
| "eval_runtime": 5768.7671, | |
| "eval_samples_per_second": 2.317, | |
| "eval_steps_per_second": 0.145, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 2.953468932850376, | |
| "grad_norm": NaN, | |
| "learning_rate": 8.025929927457941e-07, | |
| "loss": 0.0, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 2.999832822513235, | |
| "step": 26916, | |
| "total_flos": 4.3736503104621773e+18, | |
| "train_loss": 0.8739807110765321, | |
| "train_runtime": 102261.0562, | |
| "train_samples_per_second": 8.423, | |
| "train_steps_per_second": 0.263 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 26916, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 2000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.3736503104621773e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |